GAIA (Global Astrometric Interferometer for Astrophysics) Mission
Gaia (mother Earth in Greek mythology) is an ESA cornerstone space astrometric mission, part of the Horizon 2000 Plus long-term scientific program, with the goal to compile a 3D space catalog of > 1000 million stars, or roughly 1% of the stars in our home galaxy, the Milky Way. Gaia will monitor each of its target stars about 70 times to a magnitude of G=20 over a period of 5 years. It will precisely chart their positions, distances, movements, and changes in brightness. It is expected to discover hundreds of thousands of new celestial objects, such as extra-solar planets and brown dwarfs, and observe hundreds of thousands of asteroids within our own Solar System. The mission will also study about 500,000 distant quasars and will provide stringent new tests of Albert Einstein's General Theory of Relativity. 1) 2) 3) 4) 5)
Cataloguing the night sky is an essential part of astronomy. Before astronomers can investigate a celestial object, they must know where to find it. Without this knowledge, astronomers would wander helplessly in what Galileo once termed a 'dark labyrinth'.
During the satellite's expected lifetime of five years, Gaia will observe each star about 70 times, each time recording its brightness, color and, most importantly, its position. The precise measurement of a celestial object's position is known as astrometry, and since humans first started studying the sky, astronomers have devoted much of their time to this art. However, Gaia will do so with extraordinary precision, far beyond the dreams of those ancient astronomers.
By comparing Gaia's series of precise observations, today's astronomers will soon be able to make precise measurements of the apparent movement of a star across the heavens, enabling them to determine its distance and motion through space. The resulting database will allow astronomers to trace the history of the Milky Way.
In the course of charting the sky, Gaia's highly superior instruments are expected to uncover vast numbers of previously unknown celestial objects, as well as studying normal stars. Its expected haul includes asteroids in our Solar System, icy bodies in the outer Solar System, failed stars, infant stars, planets around other stars, far-distant stellar explosions, black holes in the process of feeding and giant black holes at the centers of other galaxies.
The primary mission objectives are:
• Measure the positions and velocity of approximately one billion stars in our Galaxy
• Determine their brightness, temperature, composition and motion through space
• Create a three-dimensional map of the Galaxy.
Additional discoveries expected:
- hundreds of thousands of asteroids and comets within our Solar System
- seven thousand planets beyond our Solar System
- tens of thousands of 'failed' stars, called brown dwarfs
- twenty thousand exploding stars, called supernovae
- hundreds of thousands of distant active galaxies, called quasars.
Gaia objective is to provide a very accurate dynamical 3D map of our Galaxy by using global astrometry from space, complemented with multi-color multi-epoch photometric measurements. The aim is to produce a catalog complete for star magnitudes up to 20, which corresponds to more than one billion stars or about 1% of the stars of our Galaxy. The instrument sensitivity is such that distances beyond 20-100 kiloparsec (kpc) will be covered, therefore including the Galaxy bulge (8.5 kpc) and spiral arms. The measurements will not be limited to the Milky Way stars. These include the structure, dynamics and stellar population of the Magellanic Clouds, the space motions of Local Group Galaxies and studies of supernovae, galactic nuclei and quasars, the latter being used for materializing the inertial frame for Gaia measurements.
Figure 1: Gaia measurements objectives (image credit: ESA, Airbus Defence and Space) 6)
Background: Gaia is ESA's second space mission dedicated to astrometry. It builds on the legacy of the successful Hipparcos mission (1989-1993). 7) Like Hipparcos, Gaia's observation strategy is based on detecting stellar positions in two fields of view separated by a 'basic angle', which for Gaia is 106.5º. This strategy allows astronomers to establish a coherent reference frame over the entire sky, yielding highly accurate measurements of stellar positions.
After a detailed concept and technology study during 1998–2000, Gaia was selected as a confirmed mission within ESA's scientific program in October 2000. It was confirmed by ESA's Science Program Committee following a re-evaluation of the science program in June 2002, and reconfirmed following another re-evaluation of the program in November 2003. The project entered Phase-B2/C/D in February 2006. As of the summer 2012, Gaia is in Phase-D (Qualification and Production) and will be launched in the second half of 2013. 8) 9) 10)
• In June 2013, ESA's billion-star surveyor, Gaia, has completed final preparations in Europe and is ready to depart for its launch site in French Guiana. The Gaia spacecraft arrived in Cayenne, French Guiana, on August 23, 2013 on board the Antonov 124 aircraft.
• On Oct. 23, 2013, ESA postponed the launch of the Gaia mission. The decision was taken due to a technical issue that was identified in another satellite already in orbit. The issue concerns components used in two transponders on Gaia that generate 'timing signals' for downlinking the science telemetry. To avoid potential problems, they will be replaced.
The transponders were removed from Gaia at Kourou and returned to Europe, where the potentially faulty components were replaced and verified. After the replacements have been made, the transponders will be refitted to Gaia and a final verification test made. As a consequence of these precautionary measures, it will not be possible to launch Gaia within the window that includes the previously targeted launch date of 20 November. The next available launch window is 17 December to 5 January 2014. 11)
• Update Oct. 20, 2013: The upcoming launch manifest of Arianespace has now been established. Gaia is scheduled for launch on 20 December.
• Update Nov. 22, 2013: The checks on the Gaia satellite are proceeding well, enabling the launch to take place on December 19, 2013 (Ref. 11).
Some astrometry basics:
The precise measurement of a celestial object's position is known as astrometry, and since humans first started studying the sky, astronomers have devoted much of their time to this art. However, Gaia will do so with extraordinary precision, far beyond the dreams of those ancient astronomers (Ref. 21). 12)
By comparing Gaia's series of precise observations, today's astronomers will soon be able to make precise measurements of the apparent movement of a star across the heavens, enabling them to determine its distance and motion through space. The resulting database will allow astronomers to trace the history of the Milky Way.
In the course of charting the sky, Gaia's highly superior instruments are expected to uncover vast numbers of previously unknown celestial objects, as well as studying normal stars. Its expected haul includes asteroids in our Solar System, icy bodies in the outer Solar System, failed stars, infant stars, planets around other stars, far-distant stellar explosions, black holes in the process of feeding and giant black holes at the centers of other galaxies. Gaia will be a discovery machine.
Stars as individuals and collectives:
To understand fully the physics of a star, its distance from Earth must be known. This is more difficult than it sounds because stars are so remote. Even the closest one is 40 trillion km away, and we cannot send spacecraft out to them to measure as they go. Nor can we bounce radar signals off them, which is the method used to measure distances within the Solar System. Instead, astronomers have developed other techniques for measuring and estimating distances.
The most reliable and only direct way to measure the distance of a star is by determining its 'parallax'. By obtaining extremely precise measurements of the positions of stars, Gaia will yield the parallax for one billion stars; more than 99% of these have never had their distances measured accurately. Gaia will also deliver accurate measurements of other important stellar parameters, including the brightness, temperature, composition and mass. The observations will cover many different types of stars and many different stages of stellar evolution.
Figure 2: Distance to a star can be calculated with simple trigonometry from the measured parallax angle (1 a.u. is 1 Astronomical Unit, or 149.6 million km), image credit: ESA/Medialab
The principles of Gaia:
At its heart, Gaia is a space telescope – or rather, two space telescopes that work as one. These two telescopes use ten mirrors of various sizes and surface shapes to collect, focus and direct light to Gaia's instruments for detection. The main instrument, an astrometer, precisely determines the positions of stars in the sky, while the photometer and spectrometer spread their light out into spectra for analysis.
Gaia's telescopes point at two different portions of the sky, separated by a constant 106.5º. Each has a large primary mirror with a collecting area of about 0.7 m2. On Earth we are used to round telescope mirrors, but Gaia's will be rectangular to make the most efficient use of the limited space within the spacecraft. These are not large mirrors by modern astronomical standards, but Gaia's great advantage is that it will be observing from space, where there is no atmospheric disturbance to blur the images. A smaller telescope in space can yield more accurate results than a large telescope on Earth.
Gaia is just 3.5 m across, so three curved mirrors and three flat ones are used to focus and repeatedly fold the light beam over a total path of 35 m before the light hits the sensitive, custom-made detectors. Together, Gaia's telescopes and detectors will be powerful enough to detect stars up to 400,000 times fainter than those visible to the naked eye.
Gaia uses the global astrometry concept demonstrated by Hipparcos. The principle is to link stars with large angular distances in a network where each star is connected to a large number of other stars in every direction. The condition of closure of the network ensures the reduction of the position errors of all stars. This is achieved by the simultaneous observation of two fields of views separated by a very stable basic angle. The spacecraft is slowly rotating at a constant angular rate of 1º/min around a spin axis perpendicular to both fields of view, which describe a great circle on the sky in 6 hours. The spacecraft rotation axis makes an angle of 45º with the Sun direction (Figure 3). A slow precession around the Sun-to-Earth direction, with a 63.12 days period, enables to repeat the observation of sky objects with 86 transits on average over the 5 years of mission.
Figure 3: Illustration of the sky scanning principle (image credit: ESA)
The resulting performance will enable a breakthrough in the astrometry field, as well regarding star position and velocity performance as for the number of objects observed.
Figure 4: Gaia will improve the accuracy of astrometry measurements by several orders of magnitude compared with previous systems and observations (image credit: ESA)
Gaia is an exceptionally complex space observatory. ESA awarded Airbus Defence and Space (former Astrium SAS,Toulouse, France) the prime contract in May 2006 to develop and build the spacecraft. Together with the German and British branches of Astrium, more than 50 industrial subcontractor companies from across Europe are involved in building this discovery machine. The Gaia DPAC (Data Processing and Analysis Consortium) will process the raw data to be published in the largest stellar catalog ever made. 13) 14) 15) 16) 17) 18) 19) 20)
The Gaia spacecraft is composed of two sections: the Payload Module and the Service Module. The Payload Module is housed inside a protective dome and contains the two telescopes and the three science instruments. They are all mounted on a torus made of a ceramic material (silicon carbide). The extraordinary measurement accuracy required from Gaia calls for an extremely stable Payload Module that will barely move or deform once in space; this is achieved thanks to the extensive use of this material. 21)
Underneath the Payload Module, the Service Module contains electronic units to run the instruments, as well as the propulsion system, communications units and other essential components. These components are mounted on CFRP (CarbonFiber Reinforced Plastic) panels in a conical framework.
Finally, beneath the Service Module, a large sunshield keeps the spacecraft in shadow, maintaining the Payload Module at an almost constant temperature of around -110ºC, to allow the instruments to take their precise and sensitive readings. The sunshield measures about 10 m across, too large for the launch vehicle fairing, so it comprises a dozen folding panels that will be deployed after launch. Some of the solar array panels that are needed to generate power are fixed on the sunshield, with the rest on the bottom of the spacecraft.
The Gaia spacecraft configuration is driven by the required very high thermo-mechanical stability of the entire spacecraft. A low disturbance cold gas micro-propulsion is used for fine attitude control. The astrometric instrument is used for precise rate sensing in fine pointing operating mode.
Table 1: Parameters of the Gaia spacecraft
Figure 5: Artist's rendition of the deployed GAIA spacecraft (image credit: ESA)
SVM (Service Module):
MSM (Mechanical Service Module): The spacecraft main structure is of hexagonal conical shape. It is a sandwich panel structure with CFRP (Carbon Fiber Reinforced Plastic) face sheets, and a central cone supporting the propellant tanks. The MSM houses instruments needed for the basic control and operation of the satellite; this includes all mechanical, structural and thermal elements that support the instrument payload and spacecraft electronics. It also includes the chemical & micro propulsion systems, the deployable sunshield with solar arrays, the payload thermal tent and harness. The module consists of a central tube that is about 1.17 m long and hosts six radial panels to create a hexagonal spacecraft shape.
The service module also houses the communication subsystem, central computer and data handling subsystem, the high rate data telemetry, attitude control and star trackers. For telemetry and telecommand, low gain antenna uplink and downlink with a few kbit/s capacity are employed. The high gain antenna used for the science telemetry downlink will be used during each ground station visibility period of an average of about 8 hours per day.
Figure 6: Photo of the SVM integration (image credit: EADS Astrium)
ESM (Electrical Service Module): The ESM design is driven by the science performance (attitude control laws with the hybridization of star tracker and payload measurements, high rate data telemetry, and regulated power bus for thermal stability). It houses the AOCS units, the communication subsystem, central computer and data handling subsystem, and the power subsystem.
Figure 7: Diagram of the ESM (image credit: EADS Astrium)
AOCS (Attitude and Orbit Control Subsystem). The AOCS subsystem is characterized by:
- High precision 3-axis control
- The ASTRO (Astrometric) instrument is used for precise rate sensing during the fine pointing operational mode
- A high precision gyroscope is used for quick and efficient transitions during the fine pointing operational mode. Three FOGs (Fiber Optics Gyroscopes) use the interference of light to detect mechanical rotation. Each unit contains four closed-loop gyroscope channels to provide built-in redundancy.
- Rugged flight-proven initial acquisition and safe modes
- Three sun acquisition sensors plus one gyroscope provide spin-axis stabilization during the L2 transfer phase of the mission
- One large field of view star sensor plus use of the main instrument SM (Sky Mapper) for the 3-axis controlled operational phase.
Gaia AOCS architecture is based on a fully redundant set of equipment. Moving parts on board are strictly minimized (e.g. no reaction wheels, no mechanically steerable antenna). The data is downlinked through a novel electromagnetically steerable phased array antenna and attitude control is provided by a micro propulsion system that has its first flight use with Gaia. An atomic clock is used for precise time-stamping.
Two Autonomous Star Trackers are used in cold redundancy three FSS (Fine Sun Sensors) are used in hot redundancy through triple majority voting. Three Gyro packages provide coarse rate measurements where each gyro package comprises two fully independent co-aligned channels (i.e. a fiber optic gyroscope sensor plus associated electronics per channel), with the channels being used in cold redundancy. As mentioned above, the payload module provides very precise rate measurements when the spacecraft is operated in fine pointing science modes. 25)
For actuators, a bi-propellant Chemical Propulsion System (2 x 8 10 N thrusters used in cold redundancy) is used for orbit maintenance and attitude control in coarse AOCS modes (circa 350 kg MON + MMH).
The Micro-propulsion System (2 x 6 proportional micro-thrusters) can provide a range of 0 - 1000 µN at a resolution of 0.1 µN. The individual thrusters are driven by the micro-propulsion electronic, which is internally redundant and used in cold redundancy (circa 57 kg of GN2). The nominal science AOCS mode uses the cold gas micro-propulsion system.
C&DMS (Command & Data Management Subsystem). The C&DMS is characterized by:
- An ERC-32 based central computer and distinct input/output units for efficient software development
- Two segregated MIL-STD-1553 B data buses: one for the payload module and one for the service module
- SpaceWire data links for high-speed payload data
- FDIR architecture aiming at preserving payload integrity, with built-in autonomy for increased availability.
The PDHU (Payload Data Handling Unit) is, among other things, the 'hard-disk' of Gaia, responsible for temporary storage of science data received from the telescope before transmission back to Earth. It will receive thousands of compressed images per second from the observing system; this data will be sorted and stored. The individual star data objects will be prioritized based on the magnitude of the star. A complex file management system allows deletion of low-priority data in the event of data rates or volumes that exceed the capacity of the storage or transmission systems.
The solid-state storage subsystem of the PDHU has a capacity of 960 GB which, while not impressive by terrestrial standards, is extremely large for a space system. It uses a total of 240 SDRAM modules, each with a capacity of 4 GB, which populate six memory boards. The PDHU controller board is responsible for communication with the other spacecraft subsystems, file system management and the management of telemetry and telecommands. 26) 27)
Figure 8: The PDHU (Payload Data Handling Unit), image credit: ESA
The PDHU communicates with the gigapixel focal plane over seven redundant 40 Mbit/s SpaceWire channels to acquire the scientific data coming from the seven VPUs (Video Processing Units) of the camera. The unit's controller sorts the incoming data according to star magnitude and manages deletion of low priority data should this become necessary. It sends data for transmission to Earth under the control of the CDMU (Command and Data Management Unit). The PDHU communicates with the CDMU via a MIL-STD-1553 data bus and delivers the science data over two 10 Mbit/s PacketWire channels.- The PDHU consumes only 26 W, has a mass of 14 kg, and occupies a volume of 2.3 liter.
EPS (Electrical Power Subsystem): The spacecraft is equipped with a 12.8 m2 high-efficiency triple-junction GaAs (Gallium-Arsenide) cell solar array, of which 7.3 m2 is in the form of a fixed solar array and 5.5 m2 is covered by 6 panels mechanically linked to deployable sunshield assembly.
For the launch, the deployable sunshield is folded against the payload module. After separation from the launch vehicle, it is deployed around the fixed solar array, in the same plane. During LEOP (Launch and Early Operations Phase), power is supplied by a 60 Ah mass-efficient Lithium-ion battery.
Optimum power supply during all phases of the mission is ensured by a PCDU (Power Control and Distribution Unit) with maximum power point tracking. The PCDU performs power management by generating a 28 V primary power bus that supplies power to all spacecraft subsystems. It also controls the battery state of charge and generates pyrotechnic commands as well as heater actuation as commanded by the C&DMS (Command & Data Management Subsystem).
Figure 9: Photo of the battery (image credit: ABSL)
Propulsion: After injection into the L2 transfer orbit by the Soyuz-Fregat launcher, a chemical bi-propellant propulsion system (8 x 10 N) is used for the transfer phase. It will cover attitude acquisition, spin control, mid-course corrections, L2 orbit injection, and safe mode.
After arriving at L2, one redundant set of micro-propulsion thrusters will control the spin and precession motion of the spacecraft. Regular orbit maintenance will be performed by using the chemical propulsion thrusters. - The spacecraft uses a cold gas micropropulsion system for fine attitude control.
CPS (Chemical Propulsion Subsystem): CPS is a bi-propellant system using two tanks of Herschel/Planck heritage filled with with a total of ~400 kg of propellant featuring a blowdown ratio of 4:3. Use of monomethylhydrazine as fuel and nitrogen tetroxide as oxidizer. The 10 N thrusters are manufactured by Astrium consisting of a platinum alloy combustion chamber and nozzle that tolerates the operational temperature of 1,500°C. The thruster can be operated in a thrust range of 6 to 12.5 N with a nominal thrust of 10 N which generates a specific impulse of 291 seconds.
MPS (Micro Propulsion Subsystem): The MPS is being used for fine attitude pointing and spin rate management. A total of 12 cold gas thrusters are installed on the spacecraft being grouped in three clusters each featuring four cold gas thrusters. The thruster system uses high-pressure nitrogen propellant to provide very small impulses with a thrust range of 1 - 500 µN. The system uses two nitrogen tanks, each containing 28.5 kg of N2, stored at a pressure of 310 bar (Ref. 19). The CG-MPS (Cold Gas-Micro Propulsion Subsystem) was developed by TAS-I and Selex ES S.p.A., Italy. 28)
RF communications: All communication with the Gaia spacecraft is done using the X-band. For TT&C (Tracking Telemetry and Command), a low gain antenna uplink and downlink with a few kbit/s capacity and an omnidirectional coverage are employed. The science telemetry X-band downlink is based on a set of electronically-scanned phased array antennae accommodated on the service module bottom panel. This high gain antenna is used during each ground station visibility period of about 8 hours per day.
The X-band payload downlink rate is 10 Mbit/s from L2. To achieve this, Gaia uses a specially designed on-board phased array antenna to beam the payload data to Earth (a conventional steerable antenna would disturbed the very precise measurements).
Gaia is equipped with a total of three communication antennas – two LGAs (Low Gain Antennas) and a single X-band Medium Gain Phased Array Antenna. One LGA is located pointing in the +X direction while the other points to –X being located on the Thermal Tent and the base of the spacecraft, respectively. The two LGAs build an omni-direction communications system for housekeeping telemetry downlink and command uplink with data rates of a few kbit/s.
The MGA (Medium Gain Antenna) is located on the base of the Payload Module, protruding the DSA(Deployable Sunshield Assembly) . This directional antenna can achieve data rates of up to 10 Mbit/s for science data and telemetry downlink and telecommand reception.
Demodulation of the uplink signal is completed by the transponder units before the data flow is passed on to the CDMU (Command & Data Management Unit). The downlink data is encoded by the CDMU and modulated in X-band within the transponders before being amplified by the SSPA (Solid State Power Amplifier). The signal is combined in the phased array of the active antenna in order to orient the beam towards the Earth.
Figure 10: Photo of the phased array antenna (image credit: EADS Astrium)
CCSDS Image Data Compression ASIC: In order to transmit all the data generated on board, a particularly challenging compression factor averaging 2.8 was necessary. Unfortunately the standard suite of algorithms was not able to reach this target, because of the peculiarities of Gaia imagery, which include 'outliers', such as bright stars and planets, and which are marred by the momentary 'hot pixels' due to cosmic rays in deep space. Instead, with the support of ESA compression experts, industry developed an ad hoc solution, enabling all Gaia mission data to reach their home planet. 29)
CWICOM (CCSDS Wavelet Image COMpression ASIC) is a very high-performance image compression ASIC that implements the CCSDS 122.0 wavelet-based image compression standard, to output compressed data according to the CCSDS output source packet protocol standard. This integrated circuit was developed by Airbus DS through an ESA contract.
CWICOM offers dynamic, large compressed-rate range and high-speed image compression potentially relevant for compression of any 2D image with bi-dimensional data correlation (such as a hyperspectral data cube). Its highly optimized internal architecture allows lossless and lossy image compression at very high data rates (up to 60 Mpixels/s) without any external memory by taking advantage of its on-chip memory – almost 5 Mbit of embedded internal memory).
Figure 11: The CWICOM ASIC is a customized microchip for imaging data compression (image credit: Airbus DS, ESA)
CWICOM is implemented using the largest matrix of the Atmel ATC18RHA ASIC family, and is provided within a standard surface mount package (CQFP 256). CWICOM offers a low-power, cost-effective and highly integrated solution for any image compression application, performing CCSDS image compression treatments without requiring any external memory. The simplicity of such a standalone implementation is achieved thanks to a very efficient internal embedded memory organization that removes any need for extra memory chip procurement and the potential obsolescence threatened by being bound to a specific external memory interface.
TCS (Thermal Control Subsystem): A deployable sunshield with optimal thermoelastic behavior, made of multi-layer insulation sheets, is attached to the service module and folded against the payload module for the launch. After separation of the Gaia spacecraft from the launch vehicle, the Sun shield is deployed around the fixed solar array, in the same plane. - A thermal tent covers the payload, offering extra protection against micrometeoroids and radiation.
The very high stability thermal control is mostly passive and is achieved through optical surface reflector material, multilayer insulation sheets on the outer faces of the service module, and a black painted cavity, supplemented by heaters where required. Thermal stability is guaranteed by a constant solar aspect angle and the avoidance (as far as possible) of any equipment switch-ON/OFF cycles during nominal operation.
DSA (Deployable Sunshield Assembly): The bottom floor of the SVM is a dodecagonal-shaped panel to comply with the 12 frames of the DSA The main structure consists of carbon-fiber reinforced plastic face sheets.
DSA is folded up during launch and is deployed early in the flight. It is required to shade the payload unit and protect it from direct sunlight that could compromise instrument accuracy. Keeping the instrument at a constant temperature prevents expansion and contraction during temperature variations which would alter the instrument geometry ever so slightly with a large effect on data quality. The DSA is 10 m in diameter.
The DSA is an umbrella-type structure that consists of MLI (Multilayer Insulation) as the primary shield material and six rigid deployment booms as well as six secondary stiffeners. These booms have a single articulation on the base of the Service Module for easy deployment in the radial direction by a spring system. Spacing cables link the booms to the others to ensure a synchronized deployment sequence. The booms and strings are located on the cold side of the cover to limit thermoelastic flexing.
Attached to the DSA are six rectangular solar panels (with triple-junction solar GaAs cells) that are constantly facing the sun once the shield is deployed. They provide 1910 W of EOL (End of Life) power.
Figure 12: Photo of Gaia's DSA deployment (image credit: Astrium SAS)
Legend to Figure 12: The DSA during deployment testing at Astrium Toulouse. Since the DSA will operate in microgravity, it is not designed to support its own weight in the one-g environment at Earth's surface. During deployment testing, the DSA panels are attached to a system of support cables and counterweights that bears their weight, preventing damage and providing a realistic test environment. The flight model thermal tent is visible inside the deploying sunshield and the mechanically representative dummy payload can be seen through the aperture in the tent.
Figure 13: Photo of the Gaia SVM in the EMC chamber at Intespace, Toulouse, during launcher EMC compatibility testing (image credit: Astrium SAS)
Figure 14: Exploded view of the Gaia spacecraft (image credit: EADS Astrium)
Table 2: Summary of spacecraft subsystems (Ref. 6)
Figure 15: Alternate exploded view of the Gaia spacecraft elements (image credit: EADS Astrium)
Figure 16: The Gaia flight model spacecraft undergoing final electrical tests at Astrium Toulouse in June 2013 (image credit: EADS Astrium)
Figure 17: Photo of the Gaia spacecraft in Nov. 2013 with an Astrium AIT engineer installing the transponders at the launch site (image credit: ESA)
Figure 18: Photo of the Gaia spacecraft, tucked up inside the Soyuz fairing, ready to be mated with the Soyuz lower stages (image credit: ESA, M. Pedoussaut) 30)
Launch: The GAIA spacecraft was launched on December 19, 2013 (09:12:19 UTC) from Kourou by Arianespace, Europe's Spaceport in French Guiana. The launch vehicle was a Soyuz-STB with a Fregat-MT upper stage The launch is designated as Soyuz flight VS06. 31) 32) 33)
- About ten minutes later, after separation of the first three stages, the Fregat upper stage ignited, delivering Gaia into a temporary parking orbit at an altitude of 175 km.
- A second firing of the Fregat 11 minutes later took Gaia into its transfer orbit, followed by separation from the upper stage 42 minutes after liftoff. Ground telemetry and attitude control were established by controllers at ESOC (European Space Operation Centre) in Darmstadt, Germany, and the spacecraft began activating its systems.
- The sunshield, which keeps Gaia at its working temperature and carries solar cells to power the satellite, was deployed in a 10 minute automatic sequence, completed around 88 minutes after launch. Gaia is now en route towards L2 (Ref. 31).
Orbit: Large Lissajous orbits around L2 (Lagrangian Point 2), about 1.5 million km from Earth. L2 offers a stable thermal environment because the sunshield will protect Gaia from the Sun, Earth and Moon simultaneously, allowing the satellite to keep cool and enjoy a clear view of the Universe from the other side. In addition, L2 provides a moderate radiation environment, which benefits the longevity of the instrument detectors.
• The critical LEOP (Launch and Early Orbit Phase) will last approximately four days. In this phase, Gaia will perform the first activations – transmitter switch ON, priming of the chemical thrusters, first attitude control and finding of the sun position – followed by the sun shield deployment. Engineers on ground will perform orbit determination, then prepare and execute the critical 'Day 2' maneuver to inject Gaia into its final transfer trajectory toward the L2 Lagrange point (Ref. 70).
• LEOP will be followed by the transfer cruise phase, lasting up to 30 days, an L2 orbit injection maneuver, then the in-orbit commissioning phase, during which all operations to prepare for the routine operational phase are performed. In particular, the scientific FPA (Focal Plane Assembly) and related avionics will be thoroughly tested and calibrated. The commissioning phase is expected to last four months.
• The insertion into the final 300,000 x 200,000 km Lissajous orbit around L2 was performed one month after launch.
Figure 19: Gaia mission scenario, from launch to in-orbit operations (image credit: ESA)
• March 7, 2016: While scanning the sky to measure the position and velocity of a billion stars, ESA's Gaia satellite also records many 'guest stars' – astronomical sources that, for a short period of time, are much brighter than usual. Some of these transient objects are stars undergoing major outburst events that suddenly boost their brightness, while others are supernovae, the powerful explosions at the end of a star's life. Among these detections, it is also possible to find entire galaxies, which might occasionally become brighter due to bursts of light caused by the accreting activity of the supermassive black holes at their core. 34)
- To further understand the physical properties of these transient sources, it is crucial to observe them for a longer period of time after the first detection. For this reason, the Gaia DPAC (Data Processing and Analysis Consortium) includes a team of scientists that are responsible of scrutinizing the data daily to look for unusual sources.
- Whenever they identify a transient object, the Gaia Science Alert Team, based at the Institute of Astronomy in Cambridge, UK, announces the detection to the astronomical community so that other scientists can follow up on the source with other telescopes, on Earth and in space.
- The team started to issue science alerts in September 2014, shortly after the beginning of the mission's routine scientific operations. However, the Science Alert Team was very soon overwhelmed with the number of potential transient sources, up to a million per day, which had to be heavily filtered to distinguish real guest stars from contaminants in the data stream. This data overload was due to the fact that, at the time, scientists in DPAC were still getting themselves acquainted with the complex nature of the observations performed by Gaia and the huge amount of information they contain.
- So, after about ten months and almost three hundred science alerts issued, publication of the alerts was paused for six months, in the second half of 2015. During this time, the Science Alert Team implemented a number of upgrades in the filtering and detection algorithms that are used to find transients. Eventually, the improved pipeline was switched on again in January 2016.
- The new filtering algorithm is now tuned to mask out very dense regions of the sky, especially towards the most crowded areas along the plane of our Galaxy, the Milky Way, where the false alarm rate is still too high. By combining this mask with an improved treatment of the Gaia data, the team managed to reduce the rate of potential transients from almost one million to a few hundreds per day.
- It still takes a lot of human work to identify the actual transient – about four per day – in the total number of candidate sources, but it is definitely more manageable.
- The team is looking forward to further improving the detection algorithms, and predictions suggest they might eventually find about ten transients per day. Many of these sources are being followed up on by observational programs in both the northern and southern hemispheres.
• November 3, 2015: A local cosmic celebrity was recently pictured among the multitude of stars and Solar System bodies surveyed by ESA's Gaia satellite: Comet 67P/Churyumov–Gerasimenko, currently accompanied by another ESA spacecraft, Rosetta. 35)
- While scanning the sky to map the positions and motions of a billion stars in our Galaxy, Gaia also picks up objects much closer to home, such as asteroids and comets in the Solar System. With its ability to detect faint and moving objects, Gaia has already identified tens of thousands of asteroids since routine science operations began in July 2014, and these data will be used to determine their orbits to unprecedented accuracy.
Figure 20: Rosetta comet seen by Gaia (image credit: ESA)
Legend to Figure 20: The image shows the comet's coma and tail. The nucleus and Rosetta, which was some 300 km from the surface at the time, are both hidden in the innermost pixel. A number of background stars are also sprinkled around the image, which measures about 4.5 arcminutes across – about one-seventh of the Moon's diameter. — While this image of Rosetta's comet has mainly a symbolic value – an ESA mission, 1.5 million km from Earth, looking at a fellow science mission and its object of study, both located over 260 million km away – scientific data were also collected during this observation. In fact, besides the special observing mode used to obtain the image, the comet was also caught by the onboard detection software as a 'suspected moving object'. Over the three-second observation, it appeared to move by some 100 km with respect to the background stars, as seen from a distance of 260 million km.
Flight Operations Experience from Gaia's first 1.5 years (Oct. 2015). 36)
1) Telescope contamination by water ice: Almost immediately after the L2 insertion maneuver and upon first operation of the telescope, it was noticed that the laser light level from the BAM (Basic Angle Monitor) was dropping (the BAM contains a laser interferometer that precisely measures the basic angle between the two fields of view). In parallel, the expected magnitude of the imaged stars was checked and the same effect seen (ruling out a laser problem). It was soon understood that the root cause was water ice sublimating onto the optical surfaces. Experts at Airbus DS, SOC (Science Operations Center) and DPAC (Data Processing and Analysis Consortium) could narrow down the problem to specific mirrors in the payload and targeted re-activation of decontamination heaters (primarily around mirror M2) was able to completely clear the problem.
- The impact of ice contaminated mirrors is two-fold. Firstly the apparent magnitude of the imaged stars gradually falls, which eventually means that the faintestend stars drop below the onboard magnitude limit and fail to be recorded and downlinked. Secondly, if left over longer time scales, the point spread function is eventually degraded. For the first problem a magnitude buffer was included onboard to allow fainter stars (than magnitude 20) to also be downlinked. This was possible due to the excellent performance of the telescope and FPA that could be used below the required magnitude limit of 20.
- After this targeted decontamination had cleared the problem a full decontamination was again performed part way through the commissioning, since the ice contamination was seen to be returning and also to rule out any connection with the Straylight problem. This procedure was seen to successfully clear the contamination problem and confirmed no link to the Straylight phenomena. The effect of ice build up and decontaminating recovery operations can be seen in Figure 21, which covers the initial operations.
Figure 21: The Gaia telescope response loss due to ice build-up and recovery using decontamination operations.
- The decontamination operations impact science due to the thermal disturbance and subsequent settling time required. Hysteresis implies that small mirror refocus operations may also need to be performed. These procedures have been optimized over time (to minimize science impact) and repeated three additional times in flight. Whilst the ice contamination has returned periodically, the rate at which it returns has dramatically reduced and the time between required decontamination operations has been increasing. The last decontamination was performed in June 2015, and since this time there has been no discernible return of the ice.
2) Telescope straylight: Also noticed early in the telescope commissioning, was an excess of straylight in particular parts of the FPA that varied with the spacecraft rotation phase. This data alone was not sufficient to disentangle the origin of the straylight (e.g. from diffracted sunlight or particular bright stellar/solar system objects), so it was decided to run two dedicated in-flight tests. In these tests, the spacecraft was reprogrammed to operate away from the nominal SAA (Sun Aspect Angle) of 45º (see Figure 3); with one test performed at 42º and one at 0º. Note that these attitudes could not be used for science due to the impact on end performance, but were purely for gathering engineering diagnostics. Additionally, operating at these non-nominal attitudes necessitated dedicated procedures to be developed and executed to perform various service module programming [e.g. the PAA (Phased Array Antenna) could not be used from this attitude].
- These tests were successful in that they unambiguously demonstrated the solar origin of the straylight, since it was dramatically reduced when operating at SAA= 0º.
- The Straylight and contamination problems occurred contemporaneously. At the time it was thought the two problems could be linked and that an ice layer may have formed on the inner tent roof, creating a bright surface that boosted straylight onto the FPA. This was disproved via the inflight full decontamination performed in commissioning (see above) and also via ground laboratory tests, which showed that thin ice layers on the roof would not be able increase straylight in line with the in-flight measurements.
- A working group was setup to investigate this issue and subsequently concluded that the extra straylight was caused by fibers protruding beyond the edge of the sunshield, which boosted diffracted light from the sun. - With the straylight understood and characterized, work commenced on developing a new onboard software version for the 7 VPUs (Video Processing Units). This new software contained updates from the commissioning phase, including functionality to mitigate the effect of straylight. This new software version was developed at Airbus DS during the second half of 2014 and delivered to ESOC Q1 2015. After validation campaign and coordination with the SOC/DPAC science segment, the new software was activated onboard and has been operational since April 2015.
3) Gaia's Big Data and Open Loop Ground Automation: With the commissioning successfully completed and recoveries and updates being worked on or in place, the routine phase of the mission started in June 2014. After an initial scanning law period where EPSL (Ecliptic Pole Scanning Law) was used, the nominal scan law was activated in September 2014. Since then the Gaia mission has been steadily acquiring the huge dataset required for the mapping products.
- Early in the routine phase it became clear that Gaia was performing so well that the pre-launch data downlink estimates used to plan the ground station time were insufficient. A number of factors were at work, but the primary factor was that Gaia can reach fainter magnitudes than required.
- In order to recover this valuable extra science data, the ESOC/FCT ( Flight Control Team) quickly investigated the possibility of obtaining extra ground station time for the mission through ESA's 35 m network. Extra ground station time was possible but the increase in expected data was more than could be handled by the existing Spacecraft Controller team through shift work (i.e. the concept was for a single pass with human operator per day, with some extra shifts during galactic plane scans that occur approximately 20% of the time). Updated data estimates showed an increase by around 45%, which required 2 passes every day, with three passes during galactic plane scans.
- The operations concept was reviewed and a solution found that would allow all the data to be downlinked without increasing the team size. This concept, termed open loop ground automation, involved taking passes without human operators present via pre-planned science downlink operations. One pass per day is maintained with a human operator in control, with an additional pass dedicated purely to science downlink executed without an operator present. Science data is downlinked to the ground station and forwarded to ESOC automatically, via pre-programmed commanding sent from the mission control system to the ground stations and commanding executed from the spacecraft onboard timeline.
- Although this system works well, there is increased risk of science data loss on the downlink in comparison to a concept fully reliant on human operators who have the possibility to intervene (e.g. by stopping the science downlink in case of problems on the ground station downlink chain or weather degrading the space to ground downlink). To mitigate this risk an OBCP (On Board Control Procedure) was developed and implemented onboard. This protects the science downlink by only allowing the downlink to continue in the presence of a good uplink carrier signal detected at spacecraft level from the ground station.
- This open loop automation concept was successfully validated and implemented in Q4 2014. Becoming fully operational by November 2014. Since then, science data return has been boosted by approximately 45%.
4) Outlook: After a successful commissioning campaign the Gaia mission is performing well and on course to deliver the expected dataset for the fields of Astrometry and Astronomy as a whole. The first map releases are planned for the summer of 2016. This time is required due to the enormous data processing task that is underway at SOC and DPAC and the necessity to obtain a minimum number of transits per star (Ref. 36).
• On August 21, 2015, ESA's billion-star surveyor, Gaia, completed its first year of science observations in its main survey mode. 37)
- After launch on 19 December 2013 and a six-month long in-orbit commissioning period, the satellite started routine scientific operations on 25 July 2014. Located at the Lagrange point L2, 1.5 million km from Earth, Gaia surveys stars and many other astronomical objects as it spins, observing circular swathes of the sky. By repeatedly measuring the positions of the stars with extraordinary accuracy, Gaia can tease out their distances and motions through the Milky Way galaxy.
- For the first 28 days, Gaia operated in a special scanning mode that sampled great circles on the sky, but always including the ecliptic poles. This meant that the satellite observed the stars in those regions many times, providing an invaluable database for Gaia's initial calibration.
- At the end of that phase, on 21 August 2014, Gaia commenced its main survey operation, employing a scanning law designed to achieve the best possible coverage of the whole sky.
- Since the start of its routine phase, the satellite recorded 272 billion positional or astrometric measurements 54.4 billion brightness or photometric data points, and 5.4 billion spectra.
- The Gaia team has spent a busy year processing and analyzing these data, en route towards the development of Gaia's main scientific products, consisting of enormous public catalogues of the positions, distances, motions and other properties of more than a billion stars. Because of the immense volumes of data and their complex nature, this requires a huge effort from expert scientists and software developers distributed across Europe, combined in Gaia's DPAC (Data Processing and Analysis Consortium).
- "The past twelve months have been very intense, but we are getting to grips with the data, and are looking forward to the next four years of nominal operations," says Timo Prusti, Gaia project scientist at ESA. "We are just a year away from Gaia's first scheduled data release, an intermediate catalogue planned for the summer of 2016. With the first year of data in our hands, we are now halfway to this milestone, and we're able to present a few preliminary snapshots to show that the spacecraft is working well and that the data processing is on the right track."
- As one example of the ongoing validation, the Gaia team has been able to measure the parallax for an initial sample of two million stars.
- The nearer a star is to the Sun, the larger its parallax, and thus the parallax measured for a star can be used to determine its distance. In turn, the distance can be used to convert the apparent brightness of the star into its true brightness or 'absolute luminosity'. Astronomers plot the absolute luminosities of stars against their temperatures – which are estimated from the stars' colors – to generate a 'Hertzsprung-Russell diagram', named for the two early 20th century scientists who recognized that such a diagram could be used as a tool to understand stellar evolution.
- The graph of Figure 22 shows the absolute luminosity of almost one million stars observed by Gaia as a function of their color. The diagram is based on a combination of data from Gaia's first year of observations and earlier data from ground-and space-based telescopes. Gaia has made an average of roughly 14 measurements of each star on the sky thus far, but this is generally not enough to disentangle the parallax and proper motions. To overcome this, the scientists have combined Gaia data with positions extracted from the Tycho-2 catalog, based on data taken between 1989 and 1993 by Gaia's predecessor, the Hipparcos satellite.
- The luminosity measurements are based on data from Hipparcos and ground-based telescopes, and the color information comes also from ground-based observations. This preliminary diagram provides a taste of what the mission will deliver in the coming years. Later it will be possible to compile a 'Hertzsprung-Russell diagram' based on the Gaia data alone.
- The data points appear to populate some characteristic regions of the diagram, with most of them distributed along the diagonal running from the top left corner to the bottom right: this is called the main sequence of stars, identifying all stars that are burning hydrogen in their cores – a phase that takes up the majority of a star's lifetime. Along the main sequence, brighter and more massive stars are located towards the top left of the diagram, whereas stars with lower masses and brightnesses are found towards the lower right.
- The large clump of data points in the right half of the graph identifies red giant stars: these are evolved stars that have exhausted hydrogen in their cores. As their cores collapse under their own weight, the outer layers of these stars inflate, creating large and cool – thus red – envelopes.
Figure 22: Gaia's first Hertzsprung-Russell diagram (image credit: ESA/Gaia/DPAC/IDT/FL/DPCE/AGIS)
• July 3, 2015: The image of Figure 23, based on housekeeping data from ESA's Gaia satellite, is no ordinary depiction of the heavens. While the image portrays the outline of our Galaxy, the Milky Way, and of its neighboring Magellanic Clouds, it was obtained in a rather unusual way. 38)
- As Gaia scans the sky to measure positions and velocities of a billion stars with unprecedented accuracy, for some stars it also determines their speed across the camera's sensor. This information is used in real time by the attitude and orbit control system to ensure the satellite's orientation is maintained with the desired precision. These speed statistics are routinely sent to Earth, along with the science data, in the form of housekeeping data. They include the total number of stars, used in the attitude-control loop, that is detected every second in each of Gaia's fields of view.
- It is the latter – which is basically an indication of the density of stars across the sky – that was used to produce this uncommon visualization of the celestial sphere. Brighter regions indicate higher concentrations of stars, while darker regions correspond to patches of the sky where fewer stars are observed.
- A few globular clusters – large assemblies up to millions of stars held together by their mutual gravity – are also sprinkled around the Galactic Plane. Globular clusters, the oldest population of stars in the Galaxy, sit mainly in a spherical halo extending up to 100 ,000 light-years from the center of the Milky Way.
- The globular cluster NGC 104 is easily visible in the image, to the immediate left of the Small Magellanic Cloud. Other globular clusters are highlighted in an annotated version of this image (Figure 24).
- Interestingly, the majority of bright stars that are visible to the naked eye and that form the familiar constellations of the sky are not accounted for in this image because they are too bright to be used by Gaia's control system. Similarly, the Andromeda galaxy – the largest galactic neighbor of the Milky Way – also does not stand out here.
- Counterintuitively, while Gaia carries a billion-pixel camera, it is not a mission aimed at imaging the sky: it is making the largest, most precise 3D map of our Galaxy, providing a crucial tool for studying the formation and evolution of the Milky Way.
Figure 23: Stellar density map (image credit: ESA/Gaia – CC BY-SA 3.0 IGO, Edmund Serpell)
Legend to Figure 23: The outline of our Galaxy, the Milky Way, and of its neighboring Magellanic Clouds, in an image based on housekeeping data from ESA's Gaia satellite, indicating the total number of stars detected every second in each of the satellite's fields of view. The brighter regions indicate higher concentrations of stars, while darker regions correspond to patches of the sky where fewer stars are observed.
- The plane of the Milky Way, where most of the Galaxy's stars reside, is evidently the brightest portion of this image, running horizontally and especially bright at the center. Darker regions across this broad strip of stars, known as the Galactic Plane, correspond to dense, interstellar clouds of gas and dust that absorb starlight along the line of sight.
- The Galactic Plane is the projection on the sky of the Galactic disc, a flattened structure with a diameter of about 100, 000 light-years and a vertical height of only 1000 light-years. Beyond the plane, only a few objects are visible, most notably the Large and Small Magellanic Clouds, two dwarf galaxies orbiting the Milky Way, which stand out in the lower right part of the image. A few globular clusters – large assemblies up to millions of stars held together by their mutual gravity – are also sprinkled around the Galactic Plane.
Figure 24: Annotated stellar density map (image credit: ESA/Gaia – CC BY-SA 3.0 IGO, Edmund Serpell)
• June 30, 2015: A full-size working model of Gaia's internal systems arrived at ESA/ESOC in Darmstadt this week. The Avionics Model is mounted in a circular setup (~4 m in diameter) representing the systems on the actual satellite, now orbiting the Sun–Earth Lagrangian point L2 . With this model, the ESA flight control specialists responsible for Gaia now have access to a fully functional test bench of the inner workings of the billion-star surveyor. 39)
- The model will remain at ESOC for the rest of the mission, with the team trained to use and maintain it with the support of Airbus Defence and Space, Toulouse, the prime contractor during Gaia's development.
Figure 25: Photo of the Gaia avionics model at ESOC to be used by flight control personnel. In the image (L-R): Sonia Perez, Andreas Rudolph, Kevin Kewin, Guillermo Lorenzo (image credit: ESA/L. Guilpain - CC BY-SA IGO 3.0)
• April 15, 2015: Using telescopes on Earth, Gaia's GBOT (Ground Based Optical Tracking) campaign has been delivering daily datasets, which are used to determine the satellite's position, since launch. Making use of data that are generated for its primary purpose, the tracking of Gaia, the campaign has now found an additional science application: the detection of asteroids. Using the astrometric pipeline, several tens of known and unknown asteroids are being detected every night. 40) 41)
Figure 26: Asteroid 0012820 "Robin Williams" discovered by the Tautenburg Observatory in 1978 and re-observed by GBOT with the VST (VLT Survey Telescope) on February 13, 2015 (image credit: GBOT office, Heidelberg)
• April 9, 2015: 2015 is the International Year of Light and marks an important milestone in the history of physics with the one-hundredth anniversary of Einstein's Theory of General Relativity. Having the ability to test some of its aspects to an unprecedented accuracy, Gaia will probe the tiny deviations predicted by General Relativity in our solar system. However, the satellite will also see other evidence at work such as gravitational lensing shown in the images of the so-called Einstein Cross, discovered in 1985 (Q2237+030), and of a very similar one (HE0435-1223), discovered in 2002. 42)
- Gravitational lensing was postulated by Einstein as a consequence of light bending in a gravitational field, although it was not seen until 1979 with the observation of two identical quasars, the Twin Quasars, located in the same direction with the same redshift.
- The two images of Figure 27 show the individual measurements of two outstanding instances of gravitational lensing, each with four lensed images of a distant quasar. The data have been collected by the Gaia astrometric detector over the last few months and have been processed with the nominal pipeline, without paying particular attention to the peculiarities of the sources. In both cases, the four images, closely packed in a square of less than two arcsec side, have been recorded as four independent sources at every passage of the system in the Gaia FOVs (Fields of View). The core of the foreground galaxy is also seen near the center of the Einstein Cross and was measured as if it were a star, while it was not detected in HE0435-1223. However, we know that since the on-board detection was done successfully, more data has been sent to the ground and small field images are available, but have not yet been analyzed.
- In both systems, the distant light source is a quasar, a very compact galaxy located at a distance of about 10 billion light years (z=1.7) while the lensing galaxy is much closer to us, but still several 100 Mpc (Megaparsec) away. The quasar, the galaxy and the observer are almost perfectly aligned (within 50 marcsec), causing the quasar light ray to pass through the galactic bulge that bends the rays, eventually producing the four images when reaching the Earth.
- Using the coordinates of the quasars, the orbit of Gaia and the nominal attitude, it was possible to predict accurately (within 0.5 s) the crossing times through the Gaia FOVs and then locate the relevant observations among the 40 million similar observations acquired by Gaia on an average day. The Initial Data Treatment developed by a DPAC team in Barcelona (and run at ESAC) produces every day a very preliminary, and crude for Gaia standards, astrometry and photometry of all the sources detected on-board. These are the positions plotted in the images with the corresponding HST images in the background. The magnitudes of the individual images range from 17 to 19. As we can see, the repeated observations collected over several months are very consistent and a straight combination yields already an absolute position of each image with an accuracy of about 50 marcsec for the Einstein Cross and better for the second source. This will improve when more data are received and with the global astrometric solution and the final spacecraft attitude. By the end of 2015 there will be 9 new observations of Q2237+030 and 16 for HE0435-1223.
- This early inspection of Gaia data proves that the on-board detector performs very well and can distinguish individual faint point sources within a second of arc of each other and carry out independent measurements at nominal accuracy.
- It is expected that Gaia will observe most of the about 100 known multi-imaged lenses and discover more around known and newly discovered quasars. All the images of these sources will be combined and systematically screened over an area of a few arcseconds square to detect further small images, possibly too faint to be seen in Gaia's standard acquisition mode.
Figure 27: The Einstein Cross (left) and HE0435-1223 (right) with Gaia astrometric positions placed over HST images (image credit: ESA/Gaia/DPAC/Christine Ducourant, Jean-Francois Lecampion (LAB/Observatoire de Bordeaux), Alberto Krone-Martins (SIM/Universidade de Lisboa, LAB/Observatoire de Bordeaux), Laurent Galluccio, Francois Mignard (Observatoire de la Côte d'Azur, Nice)
Legend to Figure 27: Gaia's on-board system was able to detect four images of the distant quasar in both cases and the intervening lens at the middle of the Einstein Cross. The positions are supplied by the Gaia Initial Data Treatment in a routine mode, with a very preliminary attitude determination. The magnitude of the images ranges from 17 to 19 and the astrometric accuracy of each position in this preliminary reduction is around 100 marcsec. It will be much improved during the global astrometric processing where the spacecraft attitude will also be solved together with the source astrometry.
• March 17, 2015: Gaia's FPA (Focal Plane Assembly) with its 106 CCDs is the biggest radiation monitor in space. The interactions with space radiation are detected on the CCDs in form of: 43)
5) Permanent performance degradations caused by:
- Flatband voltage shifts as result of the Total Ionizing Dose (TID)
- Increased CTI (Charge Transfer Inefficiency) as result of newly created charge trapping sites through displacement damage under NIEL (Non-Ionizing Energy Loss).
6) Transient effects leaving charge cloud tracks in the CCD pixels, when high-energy particles create electron-hole pairs along their trajectory through the silicon lattice. Different particles can leave these tracks, but for simplification the appearance of the deposited signal track on the CCD images will be called "Cosmic Rays" (CR) throughout this document.
While the permanent performance degradations are subject to other studies, the interest for this data release is the detection of cosmic rays. Gaia has no shutter and therefore no dark images are taken to be able to assess the radiation background without any other external signal sources. Furthermore nearly all CCDs are read out in windowing mode with the windows centered on the on-board detected objects, which makes cosmic ray analysis on the received science data a difficult task.
The high volume of object detections and the limited transmission bandwidth between L2 and Earth lead to optimization of the generated on-board data volume and due to Gaia's self-calibrating scheme, only few additional calibration data needs to be acquired in addition to the science objects.
One of the most suitable sets of data items for cosmic rays is the occasional engineering data from the readout process of the BAM (Basic Angle Monitor) CCDs acquired through the on-board SIF (Service Interface Function), which allows downlinking information otherwise discarded onboard before transmission. The BAM is Gaia's metrology system to measure the variations of the angle between both telescopes' line-of-sights based on laser interferometry. To be able to expose the static laser interference pattern, these are the only CCDs on board Gaia, which are operated in a stare mode with several seconds exposure time in addition to the TDI readout mode.
The BAM readout windows are centered on the highest signal interference fringes and therefore unsuitable for cosmic ray analysis. In regular intervals throughout the mission an extended length readout strip with the same BAM window width and location is recorded through the SIF mode. These window extensions contain a substantially reduced signal, which in the extreme areas of the window is completely dominated by the TDI transfer signal integration and not by exposure. Cutouts of these extreme areas are pre-processed for the purpose of cosmic ray assessment. This technical note describes the data format and content of the images.
The two BAM CCDs are located at one corner of the Gaia FPA (Figure 28, bottom left). They are divided into a nominal and redundant system marked BAM_N and BAM_R, respectively, both fed from its independent laser interferometry system. Only one system is working at a given time, and redundancy is meant here as duplication with the best performing system working during operations.
Apart from accumulated ionizing dose and proton fluency over the mission lifetime, transient events in form of cosmic ray background can give an indication of the radiation environment at Gaia's operational orbit around the Sun-Earth Lagrange point L2. Cosmic ray transient events on CCDs leave a characteristic charge trail in the images, which can be analyzed in terms of deposited charge, angle of incidence and track length. To support these studies, a small data set is being made available here (Ref. 43) together with a technical note that describes the data release content and format in more detail.
The 152 Flexible Image Transport System (FITS) image files are subsets of the occasional engineering data from the readout process of the BAM CCDs acquired through the on-board SIF (Service Interface Function), which allows downlinking information otherwise discarded on-board before transmission. These BAM CCD images provide the most suitable data set, because of two reasons: (1) these are the only CCDs on board Gaia, which are operated in a stare mode with several seconds exposure time in addition to the nominal TDI readout mode, and (2) the particular readout window subsets contain apart from the cosmic rays no other external signal source but a relatively low and constant background that has been removed in a pre-processing step.
The complete data set covering data acquisitions from May 2014 to January 2015 and the corresponding technical note are available for download from the links of Ref. 43); the data files are also in Ref. 43).
Figure 28: Location of the two BAM CCDs marked "BAM_N" and "BAM_R" in the context of the Gaia FPA (image credit: ESA/ESAC)
• January 2015: Routine operations continue. On average some 40 million stars cross the focal plane every day triggering astrometric, photometric and spectroscopic measurements. Owing to the nature of the astrometric measurements, no early data release is possible, but science alerts are issued regularly. In the meantime, the user community has been kept up to date with news items and data demonstrating the performance of Gaia. 44)
• January 2015: One year ago on January 8, 2014, the Gaia spacecraft had reached its operational orbit at L2 (launch Dec. 19, 2013). Nominal scanning started in September 2014. At this point, Gaia was working so well that it was producing more data than originally foreseen, since it was able to see stars fainter than required. Towards the end of the year, operators had to come up with a method to partially automate ground operations allowing Gaia to take advantage of more ground station time and to expand its mapping data set. 45)
- The science data are now coming down in huge quantities (11 billion camera transits were recorded by the one-year launch anniversary), with anticipation slowly building for what Gaia may find in the coming years.
• Dec. 19, 2014: Gaia was launched one year ago on Dec. 19, 2013. After an exciting year with a successful L2 orbit insertion, a challenging commissioning period and the start of routine operations, Gaia is now scanning the sky, mapping on average 40 million stars a day. Gaia's routine phase started on 25 July 2014, and the spacecraft's current state and usage of consumables indicate that the mission can be extended past its nominal 5-year lifetime.
- In the past year Gaia has recorded 11.1 billion transits with 120.5 billion astrometric, 22.2 billion photometric and 3.3 billion spectroscopic measurements with its 106 CCDs on board. In the routine phase the corresponding numbers are 6.76 billion transits with 73.4, 13.5 and 2 billion astrometric, photometric and spectroscopic measurements. The total data volume from Gaia so far is 9.2 TB. The whole sky has been observed at least once in the routine phase. The spacecraft is currently performing astrometry and photometry for stars brighter than G = 20.7 mag and spectroscopy till GRVS = 16.2 mag.
• September 2014: While scanning the sky to measure the positions and movements of stars in our Galaxy, Gaia has discovered its first stellar explosion in another galaxy far, far away (Figures 29 and 30). This powerful event, now named Gaia14aaa, took place in a distant galaxy some 500 million light-years away, and was revealed via a sudden rise in the galaxy's brightness between two Gaia observations separated by one month. 46)
- Gaia, which began its scientific work on 25 July, repeatedly scans the entire sky, so that each of the roughly one billion stars in the final catalogue will be examined an average of 70 times over the next five years.
- Discovery of supernova Gaia14aaa: In Figure 29, the light curve shows how the galaxy significantly brightened up between the two consecutive Gaia observations because of a stellar explosion, or supernova, which was named Gaia14aaa. This is the first supernova discovered with Gaia. 47)
Figure 29: Light curve of galaxy SDSS J132102.26+453223.8 obtained with Gaia [image credit: ESA/Gaia/DPAC/Z. Kostrzewa-Rutkowska (Warsaw University Astronomical Observatory) & G. Rixon (Institute of Astronomy, Cambridge)]
Legend to Figure 29: It shows the evolution in time of the galaxy's brightness. The brightness is indicated on the vertical axis; smaller magnitude values indicate a brighter source. The data points and error bars at the lower left corner are from the first observation, performed on 31 July 2014, and they are in line with previous observations of the same galaxy performed with other telescopes. The data points at the upper right corner are from the second observation, performed on 30 August 2014, and reveal a sudden rise in brightness of almost two magnitudes (roughly a factor of 6). - Using data from Gaia and other telescopes, astronomers confirmed that Gaia14aaa is a Type Ia supernova, the explosion of a white dwarf caused by the accretion of matter from a companion star in a binary system.
Figure 30: Artist's impression of a Type Ia supernova — the explosion of a white dwarf locked in a binary system with a companion star (image credit: ESA, ATG medialab, C. Carreau)
Legend to Figure 30: While other types of supernovas are the explosive demises of massive stars, several times more massive than the Sun, Type Ia supernovas are the end product of their less massive counterparts. Low-mass stars, with masses similar to the Sun's, end their lives gently, puffing up their outer layers and leaving behind a compact white dwarf. Due to their high density, white dwarfs can exert an intense gravitational pull on a nearby companion star, accreting mass from it until the white dwarf reaches a critical mass that then sparks a violent explosion.
• Routine operations started immediately after the commissioning period. At the beginning, the EPSL (Ecliptic Pole Scanning Law) was used, for 28 days, to guide the mapping of the sky. In this mode every scan crossed both the North and the South Ecliptic Pole. The advantage was that the stellar fields at the poles were studied in detail before the Gaia launch and could therefore be used for calibration. 48)
- During the second routine operations month, Gaia was commanded to follow the NSL (Nominal Scanning Law). The transition from EPSL to NSL was done in a continuous fashion to transfer the calibration over to the nominal mode. At the end of September, Gaia's spin axis orientation was optimized to catch a bright star close to the limb of Jupiter later in the mission for a light deflection experiment.
• August 4, 2014: Seven months after launch and following extensive commissioning, ESA's Gaia satellite is ready to start its scientific mission. The science phase formally began on July 18, 2014, meaning that Gaia is now under the responsibility of the Mission Manager, William O'Mullane, and the Science Operations Center team at ESA/ESAC (European Space Astronomy Center) in Madrid, Spain. 49)
On this occasion, the Gaia flag also changed home. The Gaia Project Manager, Giuseppe Sarri, handed it over to William O'Mullane, who is holding the flag in this photo (Figure 31, right) together with Payload Manager Philippe Garé (left). This symbolic handover marks a new chapter in the mission's story.
This is one of two flags that flew at Europe's Spaceport in Kourou, French Guiana during the weeks leading up to the launch on December 19, 2013. Shortly after, the other flag was sent to ESA/ESOC (European Space Operations Center) in Darmstadt, Germany, where the Mission Operations Center has since been taking care of Gaia's health.
While the Mission Operations Center receives the data from the satellite, the Science Operations Center coordinates their distribution to the Data Processing and Analysis Consortium, who will eventually produce the Gaia catalog.
The design on the flag represents the attempt of humankind to reach for the stars, an enduring fascination that resonates with the scientific goals of the Gaia mission. Originally designed for the launcher fairing, the logo will now fly at two of ESA's establishments in Europe, looking forward to Gaia's measurements and to the many discoveries to come.
Figure 31: Handover of the Gaia flag by ESA Teams to mark the start of science operations (image credit: ESA)
• July 29, 2014: The Gaia mission is operational to begin its five-year science phase. Following extensive in-orbit commissioning and several unexpected challenges, ESA's billion-star surveyor, Gaia, is now ready to begin its science mission. 50)
The commissioning phase uncovered some unexpected anomalies:
- One problem detected early in the commissioning was associated with water freezing on some parts of the optics, causing a temporary reduction in transmission of the telescopes. This water was likely trapped in the spacecraft before launch and emerged once it was in a vacuum. Heating the affected optics to remove the ice has now largely solved this problem, but it is likely that one or two more 'decontamination' cycles will be required during the mission to keep it in check.
- Another problem is associated with 'stray light' reaching Gaia's focal plane at a level higher than predicted before launch. This appears to be a mixture of light from the Sun finding its way past Gaia's 10 m diameter sunshield and light from other astronomical objects, both making their way to the focal plane as a diffuse background.
The commissioning of Gaia came to its formal end on July 18, 2014 when the board members of the MIOCR (Mission In-Orbit Commissioning Review) confirmed the readiness of the space and ground segments to start routine operations. The review summarized the commissioning activities both on ground and in orbit. New scientific performance estimates have been calculated since using in-orbit commissioning data. 51)
On July 25, 2014, Gaia started its routine phase by scanning the sky for 28 days using the so-called ecliptic-poles scanning law. This is useful to bootstrap the basic calibrations of the data. After these 28 days, the nominal scanning law will be used to determine how Gaia is scanning the sky. Although the commissioning phase has ended, some activities remain to be completed. The root causes of the stray light and the basic-angle variations have not been found yet. A dedicated working group will hence address these topics during the remainder of 2014. Nonetheless, Gaia has started its 5-year journey today to produce a map of the Galaxy in three dimensions for one billion stars with unprecedented precision!
• June 16, 2014: A series of exhaustive tests have been conducted over the past few months to characterize some anomalies that have been revealed during the commissioning of Gaia. Key among these are an increased background seen in Gaia's focal plane assembly due to stray light entering the satellite and reduced transmission of the telescope optics. In an effort to understand both problems, much of the diagnostic work has been focussed on contamination due to small amounts of water trapped in the spacecraft before launch that has been "outgassing" now that Gaia is in a vacuum. The Gaia payload module is illustrated in Figure 34. 52)
The water vapor freezes out as ice on cold surfaces and since Gaia's payload sits at temperatures between –100 and –150°C in the dark behind the big sunshield, that is where it ends up, including on the telescope mirrors. The ice initially led to a significant decrease in the overall transmission of the optics, but this problem was successfully dealt with by using heaters on Gaia's mirrors and focal plane to remove the ice, before letting them cool down to operational temperatures again.
Some ice on the mirrors was expected – that is why the mirrors are equipped with heaters – but the amount detected was higher than expected. As the spacecraft continues to outgas for a while, future 'decontamination' campaigns are foreseen to keep the transmission issue in check using a much lighter heating procedure to minimize any disturbing effect on the thermal stability of the spacecraft.
With regards to the stray light, our analysis of the test data indicates that it is a mixture of sunlight diffracting over the edge of the sunshield and brighter sources in the 'night sky' on the payload side, both being scattered into the focal plane. A model has been developed which goes some way to explaining the stray light seen in the focal plane, but not all aspects are yet understood.
One key working hypothesis was that ice deposits have built up on the ceiling of the thermal tent structure surrounding the payload, and that scattering off this ice might enhance the stray light. Although there is no way to directly confirm that this is indeed the situation, the Gaia project team nevertheless considered ways of removing any such ice.
Unlike the mirrors and focal plane, the thermal tent does not have any heaters, so alternative solutions had to be explored. One option analyzed in detail would involve altering the attitude of the spacecraft to allow sunlight to directly enter the thermal tent in order to remove any ice that might be there. The risks associated with this concept were assessed, and software and procedures developed to carry it out safely, but there is currently no plan to do so.
Under the assumption that the stray light cannot be completely eliminated, the team is investigating a variety of modified observing strategies to help reduce its impact over the course of the mission, along with modified on-board and ground software to best optimize the data that will be collected. Even if the team has to work with the stray light, it is already known that it will only affect the quality of the data collected for the faintest of Gaia's one billion stars.
Stray light increases the background detected by Gaia and thus the associated noise. The impact is largest for the faintest stars, where the noise associated with the stellar light itself is comparable to that from the background, but there is minimal impact on brighter ones, for which the background is an insignificant fraction of the total flux.
The stray light is variable across Gaia's focal plane and variable with time, and has a different effect on each of Gaia's science instruments and the corresponding science goals. Thus, it is not easy to characterize its impact in a simple way.
However, broadly speaking, the team's current analysis is that if the stray light remains as it is today, its impact will be to degrade the astrometric accuracy of a solar-type star at magnitude 20, the faint limit of Gaia, by roughly 50%, from 290 µarcsec to 430 µarcsec by the end of the mission. Things improve as one moves to progressively brighter stars, and by magnitude 15, the accuracy will remain unaltered at approximately 25 µarcsec.
It is important to realize that for many of Gaia's science goals, it is these relatively brighter stars and their much higher accuracy positions that are critical, and so it is good to see that they are essentially unaffected. Also, the total number of stars detected and measured will remain unchanged.
For brightness and low-resolution spectroscopic measurements made by Gaia's photometric instruments, current indications are that the faintest stars at magnitude 20 will have been measured to roughly the 6–8% level by the end of the mission, rather than a nominal 4%, while brighter stars will remain more accurate at about 0.4%.
The RVS (Radial Velocity Spectrometer) is most affected by the stray light and about 1.5 magnitudes of sensitivity could be lost, although the number of stars that that translates into will not be known until on-going data analysis is complete.
Finally, Gaia also contains a laser interferometer called the BAM (Basic Angle Monitor) system, designed to measure the angle of separation between Gaia's two telescopes to an accuracy of 5 µarcsec every few minutes. This is necessary in order to correct for variations in the separation angle caused by 'normal' thermal changes in the payload as Gaia spins. The system is working as planned, but is seeing larger-than-expected variations in the basic angle. The team is currently examining these data to discover if this issue will have any impact.
A comprehensive understanding of these issues will be given, when a thorough analysis of all engineering tests is complete. Gaia has nearly completed its performance verification data taking, and is about to start a month-long dedicated science observation run. Once the data have been fully analyzed, the team will be able to provide a detailed quantitative assessment of the scientific performance of Gaia (Ref. 52).
• February 2014: The Gaia observatory is slowly being brought into focus. A test calibration image (Figure 32), taken as part of commissioning the mission to 'fine tune' the behavior of the instruments, is one of the first proper 'images' to be seen from Gaia, but ironically, it will also be one of the last, as Gaia's main scientific operational mode does not involve sending full images back to Earth. — Once Gaia starts making routine measurements, it will generate truly enormous amounts of data. To maximize the key science of the mission, only small 'cut-outs' centered on each of the stars it detects will be sent back to Earth for analysis. 53)
In the commissioning phase, the telescopes must be aligned and focused, along with precise calibration of the instruments, a painstaking procedure that will take several months — to understand the full behavior and performance of the instruments — before Gaia is ready to enter its five-year operational phase. As part of that process, the Gaia team has been using a test mode to download sections of data from the camera, including the image of NGC1818 (Figure 32), a young star cluster in the Large Magellanic Cloud. The image covers an area less than 1% of the full Gaia field of view.
Figure 32: Gaia calibration image shows a dense cluster of stars in the Large Magellanic Cloud, a satellite galaxy of our Milky Way (image credit: ESA, DPAC, Airbus DS)
• With a final, modest, thruster burn on January 14, 2014, ESA's billion-star surveyor finalized its entry into its orbit around 'L2', a virtual point far out in space. L2 provides a moderate radiation environment, which helps extend the life of the instrument detectors in space. However, orbits around L2 are fundamentally unstable. 54)
- Lissajous orbit: In terms of the math, the thruster burns on January 2014 are moving Gaia onto what's known as a 'stable manifold' – a pathway in space that will lead the spacecraft to orbit around L2. Gaia is now moving in a so-called Lissajous orbit around L2, once every 180 days. - The name Lissajous refers to the shape of the path traced out by the orbit as seen from Earth, which will rise then fall above and below the ecliptic plane (the plane of Earth's orbit around the Sun) while sometimes leading and sometimes lagging the Earth. 55) 56)
Figure 33: Schematic view of Gaia's Lissajous orbit about L2 (image credit: ESA)
• January 08, 2014: The Gaia spacecraft is now in its operational orbit around the Lagrangian point L2, a gravitationally stable virtual region in space, 1.5 million km from Earth. 57)
- Entering orbit around L2 is a rather complex endeavor, achieved by firing Gaia's thrusters in such as way as to push the spacecraft in the desired direction whilst keeping the Sun away from the delicate science instruments.
- Once the spacecraft instruments have been fully tested and calibrated – an activity that started en route to L2 and will continue for another four months – Gaia will be ready to enter a five-year operational phase.
• Dec. 20, 2013: Gaia performed an important thruster burn to set course to its destination. The critical maneuver boosts Gaia into its 263,000 x 707,000 x 370,000 km, 180 day-long orbit around L2.
The first 2 days of operations were critical for the mission success. After separation from the launcher, the spacecraft starts an autonomous sequence that consisted in the main following steps:
- Telemetry initialization
- Payload bipods release
- Chemical propulsion system activation
- Sunshield (DSA) deployment
- Second sun acquisition.
PLM (Payload Module):
The payload module is housed inside a geometrical, dome-like structure called the 'Thermal Tent' (Figures 14 and 15). The payload consists of a single integrated instrument (Figure 34) that comprises three major functions. In the earlier spacecraft designs, the three functions were distributed over three separate instruments. Now the three functions are built into a single instrument by using common telescopes and a shared focal plane: 58) 59) 60) 61) 62)
7) The Astrometric instrument (ASTRO) is devoted to star angular position measurements, providing the five astrometric parameters:
- Star position (2 angles)
- Proper motion (2 time derivatives of position)
- Parallax (distance)
ASTRO is functionally equivalent to the main Hipparcos instrument.
8) The Photometric instrument provides continuous star spectra for astrophysis in the band 320-1000 nm and the ASTRO chromaticity calibration
9) The RVS (Radial Velocity Spectrometer) provides radial velocity and high resolution spectral data in the narrow band 847-874 nm.
Each function is achieved within a dedicated area on the focal plane. Afocal elements are located close to the focal plane for the photometric and spectroscopic functions, providing dispersion of the star's spectrum along the scan. This allows both functions to take benefit from the two viewing directions and from the large ASTRO aperture, and to operate in densely populated sky areas. RVS is implemented as a grating plate, combined with four prismatic spherical lenses. This allows the necessary dispersion value to be met while correcting most of the telescope aberrations. 63)
Figure 34: Annotated diagram of the Payload Module (image credit: ESA)
Legend to Figure 34: The focal plane is hanging on the 'optical bench torus' made of silicon carbide. The optics consist of 10 mirrors and the refractive optical elements. Mirrors M1, M2 and M3 form one telescope and M1', M2', M3' the other telescope. The subsequent set of mirrors M4/M4', M5 and M6 combine the light from both telescopes and direct it to the focal plane assembly. The fields of view of the two telescopes are 106.5º apart (Astrium SAS).
The payload design is characterized by:
• A dual telescope concept, with a common structure and a common focal plane. Both telescopes are based on a TMA (Three Mirror Anastigmat) design. The beam combination is achieved in image space with a small beam combiner, rather than in object space as was done in the Hipparcos satellite. This saves the mass of the beam combiner, simplifies the accommodation and eliminates the directional ambiguity of the detected targets.
• The use of SiC (Silicon Carbide) ultra-stable material for mirrors and telescope structure provides low mass, isotropy, thermo-elastic stability and dimensional stability in a space environment. This allows to meet the stability requirements for the basic angle between the two telescopes with a passive thermal control instead of an active one.
• A highly robust BAM (Basic Angle Measurement) system.
• A large common focal plane shared by all instruments.
Gaia contains two identical telescopes, pointing in two directions separated by a 106.5º basic angle and merged into a common path at the exit pupil. The optical path of both telescopes is composed of six reflectors (M1-M6), the last two of which are common (M5-M6). Both telescopes have an aperture of 1.45 m x 0.5 m and a focal length of 35 m. The telescope elements are built around the hexagonal optical bench with a ~3 m diameter, which provides the structural support.
Figure 35: Diagram of the hexagonal optical bench and the mirror system, together with the focal plane (EADS Astrium)
Although the optical design is fully reflective, based on mirrors only, diffraction effects with residual aberrations induce systematic chromatic shifts of the diffraction images and thus of the measured star positions. This effect, usually neglected in optical systems, is also critical for Gaia. These systematic chromatic displacements will be calibrated as part of the on-ground data analysis using the color information provided by the photometry of each observed object.
The main objective of the astrometric instrument (ASTRO) is to obtain accurate measurements of the relative positions of all objects that cross the fields of view of Gaia's two telescopes. The two fields of view are combined onto the single focal plane.
During its five-year mission, Gaia will systematically scan the whole sky and will have obtained some 70 sets of relative position measurements for each star. These permit a complete determination of each star's five basic astrometric parameters: two specifying the angular position, two specifying the proper motion, and one - the parallax - specifying the star's distance. The five-year long mission also permits the determination of additional parameters, for example those relevant to orbital binaries, extra-solar planets, and solar-system objects.
By measuring the instantaneous image centroids from the data sent to ground, Gaia measures the relative separations of the thousands of stars simultaneously present in the combined two fields. The spacecraft operates in a continuously scanning motion, such that a constant stream of relative angular measurements is built up as the fields of view sweep across the sky. High angular resolution (and hence high positional precision) in the scanning direction is provided by the large primary mirror of each telescope. The wide-angle measurements provide high rigidity of the resulting reference system.
Design: The astrometric instrument (ASTRO) comprises the two telescopes and the dedicated area of 62 CCDs in the focal plane, where the two fields of view are combined onto the AF (Astrometic Field). Each CCD is read out in TDI (Time Delay Integration) mode, synchronized to the scanning motion of the satellite. In practice, stars entering the combined field of view first pass across the column of the SM (Sky Mapper) CCDs, where each object is detected. Information on an object's position and brightness is processed on board in real-time to define the windowed region around the object to be read out by the following CCDs.
On-ground Data Processing: The a posteriori on-ground data processing is a highly complex task, linking all relative measurements and transforming the location (centroiding) measurements in pixel coordinates to angular field coordinates through a geometrical calibration of the focal plane, and subsequently to coordinates on the sky through calibrations of the instrument attitude and basic angle.
Further necessary corrections to be performed include those for optics effects (systematic chromatic shifts and aberration) and general-relativistic effects (light bending due to the Sun, the major planets plus some of their moons, and the most massive asteroids).
Accuracy: The accuracy of the measurements depends on the stellar type and relies on the stability of the basic angle of 106.5° between the two telescopes. This angle is monitored by the BAM (Basic Angle Monitoring) system.
The photometer will measure the SED (Spectral Energy Distribution) of all the detected objects. This will serve two goals:
• From the SED measurements, astrophysical quantities such as luminosity, effective temperature, mass, age, and chemical composition are derived
• In order to meet the astrometric performance requirements, the measured centroid positions must be corrected for systematic chromatic shifts induced by the optical system. This is only possible with the knowledge of the spectral energy distribution of each observed target in the wavelength range covered by the CCDs of the main astrometric field (~320-1000 nm).
Design: The photometer (like the spectrometer) is merged with the astrometric function, using the same large collecting apertures of the two telescopes. The photometry function is achieved by means of two low dispersion optics located in the common path of the two telescopes: one for the short wavelengths (BP) and one for the long wavelengths (RP).
- BP (Blue Photometer): Measurement in the 320-660 nm spectral range
- RP (Red Photometer): Measurement in the 650-1000 nm spectral range.
The baseline design uses only one prismatic element in fused silica for each photometer, to disperse the collected light along scan prior to detection.
The prisms are located at the nearest possible position from the focal plane, in order to facilitate the mechanical holding and moreover reduce the shadowed areas. Both prisms are attached to the box shaped CCD radiator directly in front of the detector array. Both photometers, BP and RP, have a dedicated CCD strip that covers the full astrometric field of view in the across-scan direction.
Accuracy: The spectral resolution is a function of wavelength as a result of the natural dispersion curve of fused silica; the dispersion is higher at short wavelengths.
The BP and RP dispersers will be designed in such a way that BP and RP spectra have similar sizes (on the order of 30 pixels along scan). BP and RP spectra will be binned on-chip in the across-scan direction; no along-scan binning is foreseen. For bright stars, single-pixel-resolution windows are foreseen to be used, in combination with TDI gates.
The end of mission sky averaged magnitude standard error will depend on the star type, magnitude and wavelength band and it will typically be in the range of 10-200 x 10-3 in magnitude.
Objectives: The primary objective of Gaia's RVS (Radial Velocity Spectrometer) instrument is the acquisition of radial velocities. These LOS (Line-of-Sight) velocities complement the proper-motion measurements provided by the astrometric instrument. To this end, the instrument will obtain spectra in the narrow near infrared band (847-874 nm) with a spectral resolution λ/Δλ of ~ 11,500.
The RVS wavelength range, 847-874 nm, has been selected to coincide with the energy-distribution peaks of G- and K-type stars which are the most abundant RVS targets. For these late-type stars, the RVS wavelength interval displays, besides numerous weak lines mainly due to Fe, Si, and Mg, three strong ionized calcium lines (at around 849.8, 854.2, and 855.2 nm). The lines in this triplet allow radial velocities to be derived, even at modest SNRs (Signal-to-Noise Ratios). In early-type stars, the RVS spectra may contain weak lines such as Ca II, He I, He II, and N I, although they will generally be dominated by Hydrogen Paschen lines.
Design and operations: The RVS instrument is a near-infrared, medium-resolution, integral-field spectrograph dispersing all the light entering the field of view. It is integrated with the astrometric and photometric functions and uses the common two telescopes.
- Wavelength range: 847 - 874 nm
- Resolution (R=λ/Δλ): ~ 11,500.
The RVS uses the SM (Sky Mapper) function for object detection and confirmation. Objects will be selected for RVS observations, based on measurements made slightly earlier in the RP (Red Photometer). Light from objects coming from the two viewing directions of the two telescopes is superimposed on the RVS CCDs.
Figure 36: Location of the RVS optical module and detectors (image credit: EADS Astrium)
The spectral dispersion of objects in the field of view is achieved by means of an optical module physically located between the last telescope mirror (M6) and the focal plane. This module contains a grating plate and four dioptric, prismatic, spherical, fused-silica lenses which correct the main aberrations of the off-axis field of the telescope. The RVS module has unit magnification, which means that the effective focal length of the RVS equals 35 m.
Spectral dispersion is oriented in the along-scan direction. A dedicated passband filter restricts the throughput of the RVS to the desired wavelength range.
The RVS-part of the Gaia FPA (Focal Plane Assembly) contains 3 CCD strips and 4 CCD rows. Each source will typically be observed during ~40 FOV transits throughout the 5-year mission. On the sky, the RVS CCD rows are aligned with the astrometric and photometric CCD rows; the resulting semi-simultaneity of the astrometric, photometric, and spectroscopic transit data will be advantageous for variability analyses, scientific alerts, spectroscopic binaries, etc. All RVS CCDs are operated in TDI (Time Delay Integration) mode.
The RVS spectra will be binned on-chip in the across-scan direction. All single-CCD spectra are foreseen to be transmitted to the ground without any further on-board (pre-)processing. For bright stars, single-pixel-resolution windows are foreseen to be used, possibly in combination with TDI gates. It is currently foreseen that the RVS will be able to reach object densities on the sky of up to 40,000 objects/ degree2.
On-ground Data Processing: Radial velocities will be obtained by cross-correlating observed spectra with either a template or a mask. An initial estimate of the source atmospheric parameters derived from the astrometric and photometric data will be used to select the most appropriate template or mask. Iterative improvements of this procedure are foreseen. For stars brighter than ~15th magnitude, it will be possible to derive radial velocities from spectra obtained during a single field-of-view transit. For fainter stars, down to ~17th magnitude, accurate summation of the ~40 transit spectra collected during the mission will allow the determination of mean radial velocities.
Atmospheric parameters will be extracted from observed spectra by comparison of the latter to a library of reference-star spectra using, for example, minimum-distance methods, principal-component analyses, or neural-network approaches. The determination of the source parameters will also rely on the information collected by the other two instruments: astrometric data will constrain surface gravities, while photometric observations will provide information on many astrophysical parameters.
Figure 37: Photo of the RVS optical module, containing a grating plate (middle), four fused-silica prismatic lenses, as well as a bandpass-filter plate (far right), image credit: Astrium SAS
Figure 38: Illustration of the payload module (image credit: EADS Astrium)
As the spacecraft slowly rotates, the light from the celestial object (that is, the image of the object) passes across the focal plane. In this way, Gaia steadily scans the whole sky as the satellite spins and gradually precesses, with each part being observed around 70 times in the course of the operational lifetime.
The Gaia focal-plane assembly is the largest ever developed for a space application, with 106 CCDs, a total of 937 Mpixels (almost 1 Gpixel) , which are around 90% light efficient (c.f. 20% typical terrestrial camera efficiency). The FPA has a physical dimension of 1.0 m x 0.4 m (Figure 39). The focal-plane assembly is common to both telescopes and serves five main functions:
1) The WFS (Wave-Front Sensor) and basic-angle monitor, covering 2+2 CCDs: a five-degrees-of-freedom mechanism is implemented behind the M2/M2' secondary mirrors of the two telescopes for re-aligning the telescopes in orbit to cancel errors due to mirror micro-settings and gravity release. These devices are activated following the output of two Shack–Hartmann-type wave-front sensors at different positions in the focal plane. The BAM (Basic Angle Monitor) system (2 CCDs in cold redundancy) consists of a Youngs-type interferometer continuously measuring fluctuations in the basic angle between the two telescopes with a resolution of 0.5 µarcsec per 15 minutes.
2) The SM (Sky Mapper), containing 14 CCDs (seven per telescope), which autonomously detects objects down to 20th magnitude entering the fields of view and communicates details of the star transits to the subsequent CCDs.
3) The main AF (Astrometric Field), covering 62 CCDs, devoted to angular-position measurements, providing the five astrometric parameters: star position (two angles), proper motion (two time derivatives of position), and parallax (distance) of all objects down to 20th magnitude. The first strip of seven detectors (AF1) also serves the purpose of object confirmation.
4) The blue and red photometers (BP and RP), providing low-resolution, spectro-photometric measurements for each object down to 20th magnitude over the wavelength ranges 330–680 nm and 640–1050 nm, respectively. The data serves general astrophysics and enables the on-ground calibration of telescope-induced chromatic image shifts in the astrometry. The photometers contain seven CCDs each.
5) The RVS (Radial Velocity Spectrometer), covering 12 CCDs in a 3 x 4 arrangement, collecting high-resolution spectra of all objects brighter than 17th magnitude, allowing derivation of radial velocities and stellar atmospheric parameters.
Figure 39: Layout of the focal plane assembly (image credit: ESA, EADS Astrium)
Table 3: Distribution of the 106 detectors over the FPA
Observation Sequence in the Focal Plane: All CCDs, except those in the SM (Sky Mapper), are operated in windowing mode: only those parts of the CCD data stream, which contain objects of interest, are read out; remaining pixel data is flushed at high speed. The use of windowing mode reduces the readout noise to a handful of electrons while still allowing reading up to 20 objects simultaneously.
- Every object, crossing the focal plane, is first detected either by SM1 or SM2. These CCDs record, respectively, the objects only from telescope 1 or from telescope 2. This is achieved by a physical mask that is placed in each telescope's intermediate image, at M4/M41 beam-combiner level.
- Next,a surrounding window is allocated to the object, which is propagated through the following CCDs of the CCD row as the imaged object crosses the focal plane; the actual propagation uses input from the spacecraft's attitude control system, which provides the predicted position of each object in the focal plane versus time. After detection in SM, each object is confirmed by the CCD detectors in the first strip of the AF1 (Astrometric Field); this step eliminates false detections such as cosmic rays.
- The object then progressively crosses the eight next CCD strips in AF, followed by the BP, RP, and RVS detectors (the latter ones are present only for four of the seven CCD rows).
- The nominal integration time per CCD is 4.42 seconds, corresponding to 4500 pixels along scan. For bright saturating objects, the integration time in AF1-AF9, BP and RP is reduced by activating electronic TDI gates in the detector over a short period corresponding to the bright star window. The purpose of the TDI gates is to lower the effective number of pixels along scan. Twelve gates are available in the detector and allow optimizing the signal collection for bright stars at the minimum expense for faint stars.
CCD characteristics: All CCDs have the same format and are derived from e2V Technologies (UK) design and are large-area, back-illuminated, full-frame devices. They are operated in TDI (Time Delay Integration) mode with a TDI period of 982.8 µs. The focal plane is passively cooled to 170 K for reducing its sensitivity to radiation. The box shaped CCD radiator provides the radiative surface with the colder internal payload cavity (120 K) as well as CCD shielding against radiation and support for the photometer prisms.
The Gaia CCDs are fabricated in three variants, AF-, BP-, and RP-type, to optimize quantum efficiency corresponding to the different wavelength ranges of the scientific functions. The AF-type variant is built on standard silicon with broadband anti-reflection coating. It is the most abundant type in the focal plane, used for all but the photometric and spectroscopic functions. The BP-type only differs from the AF-type through the blue-enhanced backside treatment and anti-reflection coating, and it is exclusively used in BP. The RP-type is built on high-resistivity silicon with red-optimized anti-reflection coating to improve near-infrared response. It is used in RP as well as in RVS.
Table 4: Summary of CCD parameters
Figure 40: Schematic view of the FPA with the CCD array. Light from the telescopes comes from the right in this view. The electronics radiator on the left marks the outside of the spacecraft (image credit: ESA)
Payload module data handling: 66)
Time reference: The star localization measurement is performed by transit detection and time measurement, which calls for a very accurate datation of object transits. For this purpose, a CDU (Clock Distribution Unit) provides all necessary timing signals and clock functions for video sample time tagging and ground based time scale correlation. All signals are generated on the basis of the highly stable central master 10 MHz Rubidium atomic clock.
Time correlation: In addition to the classical corrected one-way path technique, it is proposed that on-board-to-UTC time correlation be performed by a specific two way process. This technique allows to cancel symmetrical delays of ionospheric or tropospheric origin, or delays of relativistic origin. This correlation performance is furthermore independent of the orbit.
On-board Payload Processing: The PDHS (Payload Data Handling System) is implemented as a set of 7 VPUs (Video Processing Units), one for each detector row of the focal plane, feeding a common 960 GB solid state mass memory at PDHU. The processing part of the PDHS has a modular architecture which follows the FPA architecture and eases the accommodation. In the case of failure of one channel, this would have little impact on the science performance. The file-organized mass memory is a standard stand-alone unit.
Although Gaia has the biggest camera that has ever flown in space, Gaia does not actually take pictures in the conventional sense. Instead, it rather tracks the stars across its sensors as the telescopes rotate and the field of view moves across the star-filled sky. In order to do so, a constant readout of the onboard CCDs is done, and this takes a lot of computing power. For this, the seven high performance VPUs (Video Processing Units) are used which interface with the 'camera'. A VPU incorporates a dedicated Astrium-developed pre-processing board, and for the bulk of the processing, a SCS750 PowerPC board from Maxwell Technologies, Inc., of San Diego, USA. Each of the VPUs exhibits a processing power of more than 1000 MIPS (>1 GIPS). A VPU has a mass of 3.2 kg and a size of 195 x 120 x 253 mm. 67) 68) 69)
On-board data processing algorithms allow computations to be made in real time, without data buffering. The hardware-software share offers full flexibility and algorithms may be modified in-flight following first in-orbit results.
A precise centroiding of bright stars is made with the measurements performed in the first 2 columns of detectors in order to determine the star velocities for monitoring the spacecraft attitude control.
Figure 41: The Gaia VPU assembly and elements - one the the 7 systems which are controlling the camera (image credit: ESA)
Figure 42: The video processing allows detecting and tracking up to 10,000 stars/s and per CCD in combined sky images of 750,000 objects/deg2 (image credit: ESA)
Metrology and alignment: Gaia embarks two specific devices, for recovery of the optical quality after launch and for continuous monitoring of the angle between the two telescope lines of sight. All these devices are operated at 130 K.
1) A five-degree-of-freedom static mechanism is implemented behind the secondary mirrors of the two telescopes , based on the TMA (Three Mirror Anastigmat) design, for securing the optical performance in orbit and cancelling static residual errors due to mirror micro-settings and gravity release. The BAM (Basic Angle Monitoring) system consists of a Fizeau interferometer, measuring fluctuations in the basic angle between the two telescopes. The image quality is measured by two Shark-Hartmann Wave Front Sensors using two dedicated CCDs in the focal plane. The signal is coming from bright stars.
2) During science operations, a Fizeau laser interferometer measures the in orbit fluctuation of the basic angle between the two input telescopes with an accuracy better than 0.5 µas. The fringe motion with respect to the detector frame provides the line-of-sight of the telescopes along scan, and therefore the basic angle variations.
ESA's most powerful ground stations, the 35 m deep-space stations in New Norcia, Australia (DSA 1), Cebreros, Spain (DSA 2), and Malargüe, Argentina (DSA 3), will be used to send commands to Gaia and receive the high volume of science data that must be returned to Earth to create Gaia's Galactic Map. During the critical LEOP phase, additional ground station support will be provided by ESA's 15 m diameter Kourou, Maspalomas and Perth stations. 70) 71)
Communications and orbit tracking:
• The end-to-end timing of the measurements must be highly precise. To this end, Gaia carries an atomic clock to time stamp the science data, which has to be matched by ultra-precise time stamping on ground at data reception. In fact, the CCSDS ground time-stamping standard that the space community uses was extended to a picosecond (10-12 s) resolution to meet the Gaia requirements, and this capability was added to ESA's Estrack Deep Space Antennas.
• The orbit of Gaia must also be determined to very high accuracy (to within 150 m at 1.5 million km). The traditional radiometric methods are supplemented by optical observations from ground-based telescopes, which take pictures of Gaia against the background stars, and ΔDOR (delta-Differential One-way Ranging) measurements in the commissioning phase (a method where multiple DSA stations are used to precisely determine the spacecraft position with respect to a Quasar).
• GBOT (Ground Based Orbit Tracking) campaign: GBOT utilizes a network of small-to-medium telescopes aiming to track the Gaia observatory. GBOT is committed to deliver one set of data per day, which allows the determination of Gaia's position good to 20 m arcsec. The GBOT data on Gaia will be included in the orbit reconstruction performed at ESOC in order to increase the accuracy of this undertaking to a level of 150 m in position and 2.5 mm/s in motion. These tight constraints are needed, to ensure that Gaia's measurements of the stars and Solar System objects are as accurate as possible. 72)
Astronomers within the DPAC community – first set up the GBOT project in early 2008 and trialled it on missions already in the same orbital location that Gaia will operate from – L2 – including NASA's WMAP and ESA's Planck satellites. This allowed the project to test their methods, and also get some clues about the probable magnitude that Gaia will have once in orbit. It is assumed that it will be around magnitude 18, but that it is still a big unknown.
Since then, a whole infrastructure was set up, developing observing techniques, a dedicated software pipeline, a database, and observatories were recruited to deliver the GBOT project data. The backbone of the data will be supplied by the 2 m Liverpool telescope, located on La Palma, Canary Islands, Spain, and the Las Cumbres Optical Global Telescope Network (LCOGT.net), which operates 1 m telescopes in Chile, South Africa, Australia and Texas. The project will also have some support from ESO's VST (2.6 m telescope at Paranal, Chile) and additional facilities will also provide data when needed.
In 2012, the project started a new fork of GBOT, radio-GBOT, which involves VLBI observations of Gaia. These are much more precise than the optical observations, but because they use more resources, the project will use this technique less often and therefore the radio data will be used only to complement the optical measurements.
The coordination of the GBOT activities is done from Heidelberg. The data reduction, analysis and storage, will be done at the Observatoire de Paris (with a mirror of the database in Heidelberg). The pipeline software, which has been developed by the GBOT group in Paris, imports and harmonizes the data obtained from the partner observatories, processes the data and finally outputs the position of Gaia. The data is then delivered to ESA's MOC (Mission Operations Center) in Darmstadt via the SOC (Science Operations Center) in Villafranca near Madrid. Likewise, the reconstructed orbit files from ESOC are retrieved by GBOT, converted into data on Gaia's position, with finder charts, and then supplied to the partner observatories.
Figure 43: Schematic view of the GBOT elements and their interrelations (image credit: ESA)
Now, shortly before the launch of Gaia, GBOT is ready for action. GBOT's observations commence about 10 days after launch; any earlier and Gaia is too bright for the instruments of the partner institutes. The project hopes to obtain motion clips of the spacecraft moving in front of star fields as the satellite journeys towards L2. It will be a challenging task for the small team, but we will do our very best to deliver!
MSC (Mission Control System):
Mission operations will be conducted by the Flight Control Team at ESOC (European Space Operations Center) in Darmstadt Germany, comprising spacecraft operations (mission planning, spacecraft monitoring and control, and all orbit and attitude determination and control) as well as scientific instrument operations (quality control and collection of the science telemetry). The ground segment at ESOC will comprise all facilities, hardware, software and documentation required to conduct mission operations.
The ground operations facilities consist of:
• Ground stations and the communications network
• Mission control center
• FCS (Flight Control System)
• Software-based spacecraft simulator
All mission and flight control facilities, except the ground stations, are located at ESOC, including the interfaces for the provision of science telemetry to the SOC (Science Operations Center) at ESA/ESAC (European Space Astronomy Center), ESA facility in Villafranca, Spain, located about 30 km west of Madrid. 73)
The science data will be distributed to ESAC after being stored in dedicated Science Data Servers at ESOC, via high-speed communication lines.
Figure 44: The Gaia Mission Control System (image credit: ESA)
The science data processing requirements for Gaia are among the most challenging of any scientific endeavor to date. Due to the immense volume of data that will be collected, for 1 billion stars, it will be a major challenge, even by the standards of computational power in the next decade, to process, manage and extract the scientific results necessary to build a 3-dimensional view of our Galaxy, the Milky Way.
A total of some 100 TB of science data will be collected during Gaia's lifetime. The estimated total data archive will surpass 1 PB (Petabyte or 1015 bytes), roughly equivalent to 1000 1 TB hard drives from a top-end home PC.
DPAC (Data Processing & Analysis Consortium):
Unlike a mission such as the Hubble Space Telescope, Gaia does not produce data that is immediately scientifically useful. The raw telemetry must first be processed before the sought after distances can be obtained, motions, and properties of the stars observed by Gaia. This immense task will be undertaken by a pan-European collaboration, the Gaia DPAC (Data Processing and Analysis Consortium). DPAC is responsible for the processing of Gaia's data with the final objective of producing the Gaia Catalogue. Drawing its membership from over 20 countries (Figure 45), the consortium brings together skills and expertise from across the continent, reflecting the international nature and cooperative spirit of ESA itself.
The DPAC consists of about 450 persons, spread over academic institutes and space agencies throughout Europe and beyond, who are actively contributing to writing the millions of lines of code needed for the data processing and to subsequently operate the software systems and validate the resulting output. Each DPAC is responsible for a different aspect of the Gaia data processing. 74) 75) 76) 77)
Figure 45: The DPAC membership map; the red dots indicate the locations of the DPCs (image credit: ESA)
Legend to Figure 45: Next to the European country DPACs, there are also members in Brazil, Canada, Chile, Israel, and the USA.
To organize the large amount of tasks to be carried out, the DPAC has been subdivided into nine specialist units known as CUs (Coordination Units). Each CU takes the responsibility for the development of a specific part of the Gaia data processing: system architecture, simulations, astrometry, photometry, spectroscopy, object processing, variability processing, astrophysical parameters, and catalog publication. The CUs draw their membership from multiple countries.
Figure 46: Schematic view of the data flow during the processing phase (image credit: ESA)
The astronomers in the CUs conceive the scientific algorithms for the data processing and also carry out a large fraction of the software development. The software is then run at one of the six DPCs (Data Processing Centers). The personnel at the data processing centers also provide the much needed software engineering expertise. Such a large software system cannot be developed and operated by astronomers alone!
The schematic of Figure 46 shows how each CU is supported by a specific DPC (indicated in red). The data exchange within DPAC will take place through the so-called MBD (Main Data Base), housed at ESAC. After the completion of a processing cycle, data is then extracted from the MDB and prepared for release.
Note that the Gaia project is unique in that the scientific data produced by DPAC are not subject to a proprietary period. On completion of a processing cycle the results are immediately available to the scientific community and also to the general public. The nine CUs and six DPCs are coordinated by an executive committee, the DPACE (DPAC Executive) as shown in Figure 47. 78)
Figure 47: The Gaia DPAC organization, CUs and DPCs (image credit: CNES)
Gaia's VO (Virtual Observatory) Big Data Archive
ESA's Gaia mission will survey the sky for at least 5 years providing high accuracy astrometry, radial velocities and multi-color photometry. The DPAC (Data Analysis and Processing Consortium) efforts will result in an astronomical catalog with unprecedented accuracy and completeness of at least 1 billion (109) sources, and over 1PB of associated data products. 79)
This brings big data challenges in storing, querying and distributing all the associated data and meta data, comparing them with other astronomical catalogues, enabling analysis, visualization, data mining and then sharing these results with other scientists. The amount of data involved forces a change of paradigm in dealing with astronomy archives. The usual usage of downloading the data to the users for her/him to work further on it needs towards evolve to a new way of working where the users' can send her/his code to the data, run it there on computing and storage services provided directly by the archive, where the data reside. The Gaia archive will provide an infrastructure to run added value interfaces and software on top of the Gaia data.
Table 5: Gaia data : events and volume
The first Gaia catalog will be publicly released to the scientific community around summer 2016, but the development of the Gaia archive has long started and an internal version is already available for the Gaia Consortium.
The term Big Data can be used and understood differently by people, but today industry widely refers to the five "Vs" (Volume, Velocity, Variety, Veracity and Value) when speaking about Big Data. From the Volume perspective, with "only" 1 PB of data produced by the end of its lifetime, Gaia could hardly be considered as a big data mission, as many other astronomy projects (in particular ground based telescope) will produce way more data volume in a shorter timescale.
The final delivery will also include all the single epoch CCD transit data that was used in the catalog computation, reaching an estimate of around 1PB at the end of the mission in 2022 (Table 5).
Nonetheless, the billions of CCD transits, measurements and spectra (Table 5) that will result into massive sources catalogs (Figure 48) for sure makes Gaia a big data challenge, from the data Velocity and Variety points of view.
Figure 48: Catalogs currently in the Gaia Archive (image credit: ESAC, DPAC)
Ensuring the Veracity of the Gaia data represents one of the main big data challenge of the Gaia data processing. And to finish, by performing astrometry, photometry and spectroscopy of about one billion objects in our Milky Way galaxy and beyond, the extent and content of the Gaia Catalogs will enable major progress to be made in many fields of galactic and stellar astronomy, hence its Value definitely places Gaia as a major Big Data project in astronomy.
Standard archive architecture:
Standard ESA space science archive architecture (Figure 3) is based on the OAIS (Open Archival Information System) architecture. Users and scientists are used to interact with data or catalogues, either through a browser user interface or scriptable command line interface. They download the data and the full catalogue, usually via FTP, to their local disk and perform their science analysis on their local computer.
This working model works fine for small amount of data, but becomes difficult as data volume grows. To reduce the data transfer burden, the users try to select region of the sky and download only part of the catalogue, and then combine it with their own datasets or catalogues already stored on their disk. This can be described as "move the data to the computing facility".
Figure 49: Standard ESA archive architecture (image credit: ESA)
The VO (Virtual Observatory) provides an unified framework which enables transparent access to astronomical science data holdings coming from various different archives. The very same command can be sent to archives located in different locations (and using their own internal storage and database systems) and present results in a consistent way to the end user or applications. Early developed interoperability VO protocols (e.g., ConeSearch, Simple Image and Simple Spectra Access) greatly facilitates access to multiple datasets, but still assumes that data are finally being downloaded to the user's computer.
Usually, science archives implements a "VO layer" on top of the existing archive infrastructure, so all the data holdings can be accessible through these VO protocols. This ensures the interoperability of the archives with other VO compliant archives and applications.
New archive paradigm for Gaia:
With the avalanche of data in astronomy, the archive model previously described reaches its limit and a new paradigm needs to be established, the so called "move the code to the data". For Gaia, the amount of data and meta data is so big that special computing infrastructure is required to efficiently handle Gaia data. For example, a query of a cone search on the ~1 billion Gaia catalog might return 10 million sources. Another typical use case is to upload a table with sources and to cross match these with the Gaia catalog. This operation is made possible with the VO TAP (Table Access Protocol), coupled with ADQL (Astronomical Data Query Language, SQL with specialized astronomical searches). TAP also supports asynchronous query, as such a cross match can require too much time to be perform interactively. UWS (Universal Worker Service) enters in action to manage these asynchronous jobs. This back-end infrastructure serves all the front-end interfaces available at the Gaia archive.
Special emphasis has been put on the database design (based on PostgreSQL and pgSphere add-on which provides spherical data types, functions, and operators for PostgreSQL) and associated indexing. Furthermore, the Gaia archive is hosted on a powerful server and the most popular catalogs are stored on PCIe SSDs disks to ensure good performance of these crossmatch functions. Some examples of time required for some of the complex crossmatches are given in Table 6.
Table 6: Crossmatch performances
Legend to Table 6: Tycho2 vs. IGSL (Initial GAIA Source List) crossmatches are even faster than the ones with 2MASS as IGSL is located in the fastest local storage (PCIe), even when IGSL (similar to the final Gaia catalog) is around 3 times bigger than 2MASS.
End user can then interact with the Gaia through a standard GUI (Graphical User Interface) from any standard web browser. This offers a full ADQL query interface, with example to help the user familiarize with the Gaia catalog content and structure.
In addition, it is expected that many of the users will interface with Gaia data directly through scriptable interface. All operations available from the Gaia archive GUI (ie TAP/ADQL queries) can also be included directly in user's scripts. Various examples of such scripts are provided in the most commonly used programming language in astronomy (Python, C, Java).
VO applications (e.g., Topcat, VOSpec) can access directly the Gaia data through the corresponding VO protocol, TAP for table, Simple Access Protocol for spectra, without the need to develop a "VO layer" as seen in the standard ESA archive architecture.
When the user retrieves big volume of Gaia data, she/he would be able to download it to her/his local disk via FTP, but it will probably be more efficient to leave it on the archive disk itself to avoid the burden of the network transfer. This can be done with VOSpace, virtual disk accessible by VO data access protocols. By keeping the data on her/his VOSpace, the user can continue to interact with it and as well share it with any other Gaia archive user, to facilitate scientific collaboration.
Figure 50: Gaia archive VO built-in architecture (image credit: DPAC)
A similar mechanism is provided for meta data storage. If the user would download the results of a complex search (i.e., returning millions of entries) on her/his computer, she/he will probably need to ingest these into a local database to continue to work on them. It appears then more efficient to provide to the user some space directly within the Gaia archive database to store the results of her/his searches. As such, the user "database" becomes VO compliant and shareable as well with any other Gaia archive user.
Overall, the Gaia archive architecture depicted in Figure 50 has been built around these VO protocols and has become the first "VO built-in" archive. By design, the Gaia archive also becomes immediately interoperable with any other VO compliant archive and application.
Gaia AVIs (Added Value Interfaces):
The user workspace can be brought one step further to enable the user to also run her/his own code directly on the Gaia archive through so called "Gaia Added Value Interfaces" (Gaia AVIs, Figure 51). Four AVI demonstrators are currently being developed for transient alerts, advanced visualization, spectral classification and temporal analysis. These AVIs will run using containers (Docker) and will make use of the Gaia Archive VO built-in protocols (TAP, VOSpace). A Gaia AVI Portal will be created and users will be able deposit their own code and can run it on their data located into their user workspace (VOSpace and database). AVI templates will be provided to help the user to develop their own AVIs.
The Gaia AVI project is currently being developed as a proof of concept project to be delivered in 2017 and could become the new framework for collaborative "Archive 2.0".
Figure 51: Gaia archive AVI building blocks (image credit: DPAC)
Big Data Visualization:
Before being able to search for Gaia data, it might be really helpful to provide visualization of the Gaia data in various ways, such as density maps, 1D histograms (Figure 52) or again interactive visualization through VO application (Aladin Lite), integrated into the archive GUI, through another VO protocol SAMP (Simple Application Messaging Protocol). The production of such graphs requires the use of big data reduction techniques, such as Map / Reduce. With 10 parallel threads on a powerful machine with big RAM (1TB), fast disks (PCIe SSDs with fast random IO) and efficient database (PostgreSQL), the production of the density maps for the GUMS simulated catalogue (2.14 billion rows) took less than 2 minutes and the ones for the IGSL (1.22 billion rows) as little as 65 seconds.
Figure 52: Gaia data density maps and 1D histograms (image credit: DPAC)
Another way to visualize Gaia data will be through the recently released science-driven discovery portal for all the ESA Astronomy Missions called "ESA Sky" that allow users to explore the multi-wavelength sky and to seamlessly retrieve science-ready data in all ESA Astronomy mission archives from a web application without prior-knowledge of any of the missions. Among other things, the system offers progressive multi-resolution all-sky projections of full mission datasets using a new generation of HiPS (Hierarchical Progressive Survey) files. HiPS is based on the HEALPix sky tessellation and is essentially a mapping of survey data at various spatial resolutions into a collection of HEALPix tiles. It is particular adapted to big data visualization as it allows a dedicated client/browser tool to access and display a survey progressively, based on the principle that "the more you zoom in on a particular area the more details show up".
Summary: Gaia, ESA's cornerstone mission currently in operations, represents one of the major big data challenge in astronomy to date. A totally new archive architecture (both hardware and software) has been developed to tackle this challenge. It results into one of the first VO built-in science archive, paving the way towards flexible, open and interoperable archive services. The user will work directly with the data in the archive through dedicated user workspace, without the need to transfer it to her/his location, She/he will be able to become an actor of the archive with the possibility to bring her/his code to the data and share it with other archive users. This new "Archive 2.0" concept will be the mean to fully exploit the science legacy of the Gaia mission.
1) "Gaia Overview," ESA, June 19, 2013, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia_overview
2) S. Clark, "GAIA - ESA's Galactic Census," ESA brochure, BR-296, June 2012, URL: http://sci.esa.int/gaia/50670-gaia-esa-s-galactic-census-esa-br-296/#
3) M.A.C. Perryman, O. Pace, "GAIA – Unravelling the Origin and Evolution of Our Galaxy," ESA Bulletin 103, August 2000, URL: http://www.esa.int/esapub/bulletin/bullet103/perryman103.pdf
4) Philippe Gare, Giuseppe Sarri, Rudolf Schmidt, "ESA's 'Billion-Pixel' Camera - The challenges of the Gaia mission," ESA Bulletin, No 137, February 2009, pp.50-59, URL: http://www.esa.int/esapub/bulletin/bulletin137/bul137g_sarri.pdf
6) Eric Maliet, Philippe Charvet, François Chassat, Giuseppe Sarri, Robert Furnell, Timo Prusti, "Gaia spacecraft in-orbit commissioning status," Proceedings of the 65th International Astronautical Congress (IAC 2014), Toronto, Canada, Sept. 29-Oct. 3, 2014, paper: IAC-14.A7.1.4
7) Note: The Hipparcos mission is named after the ancient Greek astronomer Hipparchus who, in 129 BC and using only naked eye observations and simple geometry, catalogued the relative positions of around one thousand stars.
8) M. McCaughrean (editor), "ESA's Report to the 39th COSPAR Meeting," (ESA SP-1323, June 2012), pp. 321-324, URL: http://sci.esa.int/home/50671-esa-report-to-the-39th-cospar-meeting-esa-sp-1323/#
9) "Guide to our Galaxy," ESA, Nov. 21, 2013, URL: http://spaceinvideos.esa.int/Videos/2013/11/Guide_to_our_Galaxy
10) L. Lindegren, M. A. C. Perryman, "GAIA Global Astrometric Interferometer for Astrophysics, A Concept for an ESA Cornerstone Mission," 1995, URL: http://gaia.obspm.fr/IMG/pdf/Lindegren_Perryman_1995.pdf
11) "Gaia launch postponement update," ESA, Oct. 23, 2013, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Gaia_launch_postponement_update
12) Devin Powell, "Europe's Star Power," Nature, Vol. 502, October 3, 2013, URL: http://www.nature.com/polopoly_fs/1.13859!/menu/main/topColumns/topLeft-Column/pdf/502022a.pdf
13) "Gaia Fact Sheet," ESA, June 19, 2013, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Gaia_factsheet
14) "Spacecraft development," EADS Astrium, Nov. 2007, URL: http://www.astro.lu.se/ELSA/pages/PublicDocuments/Moisson.pdf
15) S. Clark, K. Fletcher, "Gaia - ESA's Galactic Census," ESA brochure, Vol. BR-296, June 15, 2012, URL: http://sci.esa.int/gaia/50670-gaia-esa-s-galactic-census-esa-br-296/#
19) Patrick Blau, "Gaia Spacecraft Overview," Spaceflight 101, URL: http://www.spaceflight101.com/gaia-spacecraft-overview.html
20) "The Gaia Spacecraft and Instruments," ESA, URL: http://www.rssd.esa.int/SA/GAIA/docs/posters/InstrumentsNoAuthor.pdf
21) Giuseppe Sarri, Timo Prusti, Ared Schnorhk, "ESA's Billion-Star Surveyor: Gaia Ready for Launch Campaign," ESA Bulletin 155, August 2013, pp. 2-11, URL: http://esamultimedia.esa.int/multimedia/publications/ESA-Bulletin-155/
22) "Service Module," ESA, June 14, 2013, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Service_module
23) "Service Module Summary," ESA, URL: http://sci.esa.int/gaia/40130-service-module/?fbodylongid=1910
25) "Gaia Mission - General Overview," ESA, April 13, 2015, URL: http://www.sscspace.com/gaia-mission-general-overview
26) "Gaia Flight Model Payload Data Handling Unit Delivered and Integrated," ESA, May 23, 2013, URL: http://sci.esa.int/gaia/51777-09-gaia-flight-model-payload-data-handling-unit-delivered-and-integrated/
27) Christophe Delay, Stephane Humbert, "1 Tbit of data serviced by SpW," Proceedings of the 5thInternational SpaceWire Conference 2013, Gothenburg, Sweden, June 10 - 14, 2013, pp: 167-170, URL: http://2013.spacewire-conference.org/downloads/SpW2013ProceedingsFinal.pdf
28) A. Polli, G. Noci, L. Ceruti, L. Fallerini, A. Atzei, "European Cold Gas Micro Propulsion System reached TRL-9 on board GAIA S/C. Discussion on in flight performance and system flexibly and capabilities, for being implemented in other missions requiring ultra-fine satellite positioning and attitude control," Proceedings of the 65th International Astronautical Congress (IAC 2014), Toronto, Canada, Sept. 29-Oct. 3, 2014, paper: IAC-14,C4,3.3
29) Raffaele Vitulli, Enrico Magli, Sean Blair, "Putting the squeeze on : How data compression can stop space missions drowning in data," ESA Bulletin, No 158, May 2014, pp: 48-59, URL: http://www.esa.int/About_Us/ESA_Publications/ESA_Bulletin_158_May_2014
30) "Gaia secured inside fairing," ESA, Dec. 13, 2013, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Gaia_secured_inside_fairing
31) "Liftoff for ESA's billion-star surveyor," ESA PR 44-2013, December 19, 2013, URL: http://sci.esa.int/gaia/53536-esa-pr-44-2013-liftoff-for-esas-billion-star-surveyor/
32) "Next stop ..... the stars! Arianespace successfully lofts Gaia on Soyuz," Arianespace, Decmber 19, 2013, Soyuz Flight VS06, URL: http://www.arianespace.com/news-mission-update/2013/1120.asp
33) Giuseppe Sarri, Timo Prusti and Ared Schnorhk, "ESA's Billion-Star Surveyor, Gaia ready for launch campaign," ESA Bulletin, No 155, August 2013, pp. 2-11
34) "Gaia science alerts are back," ESA, March 7, 2016, URL: http://sci.esa.int/gaia/57552-gaia-science-alerts-are-back/
35) "Celebrity comet spotted among Gaia's stars," ESA, Nov. 3, 2015, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Celebrity_comet_spotted-among_Gaia_s_stars
36) D. Milligan, A. Rudolph, G. Whitehead, T. Loureiro, E. Serpell, F. di Marco, J. Marie, E. Ecale, "ESA's Billion Star Surveyor – Flight Operations Experience from Gaia's first 1.5 years," Proceedings of the 66th International Astronautical Congress (IAC 2015), Jerusalem, Israel, Oct.12-16, 2015, paper: IAC-15-B6.3.2
37) "Gaia's first year of scientific observations," ESA, August 25, 2015, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Gaia_s_first_year_of-scientific_observations
38) "Counting stars with Gaia," ESA, July 3, 2015, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Counting_stars_with_Gaia
39) "Surrounded by Gaia," ESA, June 30, 2015, URL: http://www.esa.int/spaceinimages/Images/2015/06/Surrounded_by_Gaia
43) Ralf Kohley, "Cosmic rays on BAM CCDs data delivery note," ESA/ESAC, March 17, 2015, Document: GAIA-DE-TN-ESAC-RKO-033, URL: https://gaia.esac.esa.int/archive-doc/docs/crays/GAIA-DE-TN-ESAC-RKO-033_issue_1_rev_0_(Cosmic_rays_on_BAM_CCDs_data_delivery_note).pdf
44) "Gaia," Status at end of January 2015, reported in Programs in Progress, ESA Bulletin No161, p. 69, May 11, 2015, URL: http://esamultimedia.esa.int/multimedia/publications/ESA-Bulletin-161/offline/download.pdf
45) Dave Milligan, "A year on-station for Gaia," January 14, 2015, URL: http://blogs.esa.int/rocketscience/2015/01/14/a-year-on-station-for-gaia/
46) "Gaia discovers its first supernova," ESA, Sept. 12, 2014, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Gaia_discovers_its_first-supernova
47) "Discovery of supernova Gaia14aaa," ESA, Sept. 12, 2014: URL: http://www.esa.int/spaceinimages/Images/2014/09/Discovery_of_supernova-Gaia14aaa
48) "Gaia mission status," ESA Bulletin, No 160, November 2014, p. 68
49) "Handover of the Gaia flag to mark start of science operations," ESA, August 4, 2014, URL: http://www.esa.int/spaceinimages/Images/2014/08/Handover_of_the_Gaia_flag-to_mark_start_of_science_operations
50) "Gaia: 'Go' for Science," ESA, July 29, 2014, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Gaia_Go_for_science
52) "Preliminary analysis of stray light impact and strategies," Update from the Gaia Project Team, June 16, 2014. URL: http://blogs.esa.int/gaia/2014/06/16/preliminary-analysis-of-stray-light-impact-and-strategies/
53) "Gaia comes into focus," ESA, Feb. 6, 2014, URL: http://sci.esa.int/gaia/53653-gaia-comes-into-focus/
54) "The Experts behind Gaia's arrival at nothingness," ESA, Jan. 15, 2014, URL: http://www.esa.int/Our_Activities/Operations/The_experts_behind_Gaia_s_arrival-at_nothingness
57) "Gaia enters its operational orbit," ESA, January 08,2014, URL: http://www.esa.int/Our_Activities/Space_Science/Gaia/Gaia_enters_its_operational-orbit
59) "Spacecraft & Instruments," ESA, URL: http://www.cosmos.esa.int/web/gaia/spacecraft-instruments
61) Philippe Gare, Giuseppe Sarri, Rudolf Schmidt, "ESA's 'Billion-Pixel' Camera," ESA Bulletin 137, February 2009, URL: http://www.esa.int/esapub/bulletin/bulletin137/bul137g_sarri.pdf
62) Patrick Blau, "Gaia Instrument Information," Spaceflight 101, URL: http://www.spaceflight101.com/gaia-science-instruments.html
63) "Gaia Mission - General Overview," SSC, April 13, 2015, URL: http://www.sscspace.com/gaia-mission-general-overview
68) "Gaia video processing unit test model delivered," ESA, Feb. 13, 2009, URL: http://sci.esa.int/gaia/44228-gaia-video-processing-unit-test-model-delivered/
71) David Milligan, Andreas Rudolph, Gary Whitehead, Tiago Loureiro, "Gaia Ground Segment Development and Operations Concept," AIAA SpaceOps 2010 Conference, Huntsville, AL, USA, April 25-30, 2010, Vol. 3, ISBN: 978-1-63266-002-2, URL: http://arc.aiaa.org/doi/pdf/10.2514/6.2010-2162
73) Rocio Guerra, "ESAC and the Gaia Catalogue," 2009, URL: http://www.ieec.cat/mediterrania2009/sites/default/files/proceedings/poster_Guerra.pdf
74) "Data Processing," ESA, URL: http://www.cosmos.esa.int/web/gaia/data-processing
75) Anthony G. A. Brown, "Gaia data processing," ESA, Sept. 23, 2013, URL: http://blogs.esa.int/gaia/2013/09/23/gaia-data-processing/
76) Veronique Valette, Kader Amsif, "CNES Gaia Data Processing Centre, a complex operation plan," Proceedings of SpaceOps 2012, The 12th International Conference on Space Operations, Stockholm, Sweden, June 11-15, 2012, URL: http://www.spaceops2012.org/proceedings/documents/id1291264-Paper-001.pdf
77) "List of Institutes involved in DPAC," ESA, URL: http://www.cosmos.esa.int/web/gaia/dpac/institutes
78) Veronique Valette, Benoit Frezouls, "Processing a billiom of star, an organizational challenge," SpaceOps 2014, 13th International Conference on Space Operations, Pasadena, CA, USA, May 5-9, 2014, paper: AIAA 2014-1701, URL: http://arc.aiaa.org/doi/pdf/10.2514/6.2014-1701
79) Christophe Arviset, Javier Durán, Juan González, Raúl Gutiérrez, José Hernández, Uwe Lammers, Bruno Merin, Sara Nieto, William OMullane, Jesús Salgado, Juan Carlos Segovia, "Big Data, Big Challenges and new Paradigm for the GAIA Archive," Proceedings of the 2016 Conference on Big Data from Space - BiDS'16, Santa Cruz de Tenerife, Spain, March 15-17, 2016, pp: 9-12, URL: http://publications.jrc.ec.europa.eu/repository/bitstream/JRC100655/LBNA27775ENN.pdf
The information compiled and edited in this article was provided by Herbert J. Kramer from his documentation of: "Observation of the Earth and Its Environment: Survey of Missions and Sensors" (Springer Verlag) as well as many other sources after the publication of the 4th edition in 2002. - Comments and corrections to this article are always welcome for further updates (firstname.lastname@example.org).