White Paper on a Radio Synoptic Survey Telescope (RSST) S.T. Myers (NRAO/LANL) and ... v1.0 31-Oct-2006 STM For LANL Grand Challenges v1.1 4-Jan-2007 STM For LANL Grand Challenges v2.0 26-Apr-2007 STM For NSF/DOE/NASA Astrophysics & Cosmology Communities v2.1 8-Jun-2007 STM For AUI Committee on U.S. Radio Astronomy v3.0 16-Jun-2007 STM Update v4.0 10-Sep-2007 STM For Chicago-3/NAIC Frontiers This white paper presents a vision for a "Radio Synoptic Survey Telescope" (RSST) as a next-generation radio array. The RSST would be targeted at questions at the forefront of astronomy, astrophysics and cosmology that are of great interest to the NSF, DOE and NASA scientific communities. The RSST concept came out of discussions that occurred at the Community Workshop "Building the Foundation for U.S. Astronomy at m/cm Wavelengths in 2010 and Beyond" held August 3-4, 2006 in Tucson, AZ (http://www2.naic.edu/~astro/chicago2/). During the course of this meeting it was recommended that Epoch of Reionization (EOR) and Cosmological HI studies were the two most mature scientific concepts ripe for presentation to the upcoming Decadal panel. The HI "machine" idea was encapsulated in the RSST concept, and was deemed to be of particularly wide interest to the astronomy, astrophysics, and physics communities and that further exploration of this concept be carried out. The RSST fulfills a number of the science goals for US Radio Astronomy put forth in the report of the Radio, Millimeter and Submillimeter Planning Group (RMSPG) (http://www.astro.cornell.edu/~haynes/rmspg/docs/rmspgreport.pdf). Introduction: _____________ In order to realize the ambitious goals for cosmology and to answer the fundamental questions about our Universe as set forth in nationally recognized documents such as "Connecting Quarks with the Cosmos: Eleven Science Questions for the New Century" and the Dark Energy Task Force (DETF) report, a number of large astronomical surveys are planned for the coming decade and beyond. These observations span nearly the entire observable electromagnetic spectrum from radio to gamma rays, and will employ ground based telescopes, balloon-borne instruments, and space observatories to probe the entirety of the observable Universe. Although this entire spectrum is accessible from space and to some extent from high-altitude balloons, ground-based telescopes can observe in three main windows. The radio part of the electromagnetic spectrum covers meter-waves from 10 MHz (the ionospheric cutoff) to the submillimeter just above 1 THz (the atmospheric cutoff). The Cosmic Microwave Background (CMB) is contained within this window, as are key molecular and atomic lines (such as the millimeter-wave CO transitions and the 21-cm HI line) and important continuum emission processes from thermal dust, free-free from ionized gas, and non-thermal synchrotron radiation from relativistic plasmas. This white paper focuses on the exploitation of the 21-cm wavelength HI line as a probe of the Universe from the present epoch back to a redshift of around z=2.5 (400 MHz to 1.4 GHz). Overarching Themes in Cosmic Evolution: _______________________________________ Cosmology is the study of the origin and evolution of the Universe. Over the past century, observational and theoretical advancements have led to a number of conclusions about the physical nature of our Universe, which are encapsulated in what can be referred to as the "Standard Cosmological Model" (not to be confused with the "Standard Model" of fundamental physics). The cornerstone of this standard model is that the Universe is evolving in time, with the expansion rate currently accelerating (requiring an equation of state with negative pressure, or "Dark Energy"). In the standard model, the Universe has always been expanding, albeit at variable rates depending on the equation of state at a given time. Overall, the corresponding temperature or energy scale has been decreasing with cosmic time, possibly from the Planck scale close to the time of the "Big Bang". If we can look into the past, we see the Universe has passed through many phase transitions, some of which correspond to the breaking of fundamental symmetries. These early stages leave fossil relics in the form of radiation, particles, or correlations in structure. Furthermore, the long-range gravitational forces are dominated by the Dark Energy plus effectively collisionless "Dark Matter". Baryons (and leptons) account for only a few percent of the energy density of the Universe, and thus the properties of the dark sector are of primary interest to cosmologists studying the history since nucleosynthesis. Observationally, the key measurements are of the expansion history (through its integral, the cosmic distance scale), and growth rate of structure (which is controlled by the expansion history and the equation of state). The cosmological history of the Universe is viewed by observing its state as a function of distance and thus look-back time. Due to the expansion, the emitted spectrum of objects is shifted and stretched by the "redshift", adding wavelength evolution into the mix. There are other phenomena that vary on shorter timescales than the expansion time. Many of the most distant phenomena, which are the signposts used to probe the early Universe, are highly energetic phenomena due to cosmic explosions or matter collapsed to extreme densities. Bursts, flares, supernovae and scintillation provide important clues to the nature of these rare objects and for their use as cosmological markers. What is the Radio Synoptic Survey Telescope (RSST)? ___________________________________________________ We propose to undertake a new initiative to focus on a next generation radio astronomical survey telescope, the RSST. The RSST is one of the concepts for a Square Kilometer Array (SKA) project. The SKA is an international project (http://www.skatelescope.org/) and is one of the future large projects recognized by the U.S. Decadal survey. In reality, the SKA is an umbrella for at least 3 different arrays, including a low-frequency array aimed at Epoch of Reionization (EOR) studies, a high-frequency array designed to replace the VLA, and an array targeted at the 21cm HI line. The idea of a large array aimed at the HI 21cm line in galaxies from z=0-2 (0.6-1.4GHz) is gaining momentum in the radio astronomical community, and has clear connections to optical/IR projects such as the Large Synoptic Survey Telescope (LSST, http://www.lsst.org) and the Supernova/Acceleration Probe (SNAP, http://snap.lbl.gov/). For example, the RSST could get HI redshifts for all galaxies the LSST observes out to z=2 that has Milky Way or more amounts of neutral gas. There are other important cosmological observations that this array can do, such as weak lensing. Note that this SKA concept was included in the DETF report and thus would have high impact for Dark Energy studies. The science case for the SKA is made in a large book, the chapters of which are available online (http://www.skatelescope.org/pages/page_sciencegen.htm). Of particular relevance are chapters 5 "Galaxy evolution, cosmology and dark energy with the Square Kilometre Array" (http://www.skatelescope.org/PDF/sciencebook/5Rawlings.pdf) and 9 "Cosmology with the SKA" (http://www.skatelescope.org/PDF/sciencebook/09Blake.pdf). These outline the goals for and sketch science surveys using an RSST-like concept. The RSST differs from the vanilla SKA concept in that it is a targeted experiment, much like SDSS, aimed first and foremost at being a HI cosmology machine. All parameters of its design will be set to maximize its capabilities in mapping HI to the highest feasible redshift, somewhere from z=1.5 to 2.5. More detailed design studies will need to be carried out over the next few years in order to be ready to present a white paper to the next Decadal Review committee in 2009. Besides technological advancements required in order to afford the large collecting area, maximize receiver sensitivity, and supply the computing power and throughput needed to acquire and process the data, studies will also need to be done to estimate the experimental parameters necessary to extract the cosmology, such as the required redshift depth. Who Will Build the RSST? ________________________ In the United States, there are a number of groups who have the necessary expertise to realize the RSST as a project. These include the research groups in the universities, national laboratories, and national radio astronomy observatories who have the critical expertise: - experts in the design, construction, and operation of radio interferometric arrays, - cosmologists, astrophysicists, and physicists that understand the theoretical goals of RSST science and the models that underpin our current state of knowledge, and - researchers in the frontiers of applied mathematics, statistics and computer science that can develop the next generation analysis techniques necessary to handle the tremendous rate and volume of data that RSST will produce. This community transcends the individual funding agencies of NSF, DOE and NASA, as evidenced by similar goals put forth in the various agency and inter-agency reports and decadal plans. There are natural links to other key astronomical endeavors such as the Large Synoptic Telescope program (of which LSST is a concept). Our colleagues in Canada also have capabilities and interests that are critical for the success of the RSST, and with the good working relationships between our respective national communities (in EVLA and ALMA for example) we should be able to from a North American RSST consortium partnership. Note that inclusion of our Mexican colleagues is not precluded either, as they have been active in astronomical partnerships such as EVLA and LMT. The Square Kilometer Array project has long been an interest of the European astronomical community. In fact, the European groups have assumed a leadership role in the organization of SKA planning activities, partly due to the slow pace of U.S. SKA related research and development this past decade. The International SKA Organization is headquartered in the Netherlands and has been extremely active in pushing forward a number of design and development related activities. Other partners, such as Australia and South Africa, have also been extremely active in the SKA. It is absolutely clear that the next generation radio array will be a large international project --- larger even than ALMA. By consolidating our own national scientific and technical resources into a project such as RSST, we can bring a strong vision backed by a strong North American partnership to the table during the critical international planning and project establishment negotiations. Where Will RSST Be? ___________________ An independent International Square Kilometer Array Steering Committee (ISSC) was established by the international organization to promote, oversee, and coordinate planning for the SKA project. This included the process for the solicitation, submission and review of proposals for the siting of the SKA. Four candidate sites were submitted for consideration in 2005: Argentina/Brazil, Australia, China, and Southern Africa. The Australian and Southern Africa locations were chosen as the short-list in September 2006. The United States SKA Consortium chose not to submit a proposal for a US-based site. There were a number of reasons for this, including the lack of large radio-quiet environments. In the context of an RSST, it may be possible to revisit this question, although it is unlikely that better sites than the two chosen by the ISSC will be found within North America. Like other research areas such as High Energy Particle Physics, the physical location of a large facility for a large international project is less important for the health of the national communities that make up the partnership than the breadth and depth of participation in the intellectual activities of that project. We now have considerable experience with successful large projects based on foreign soil, such as ALMA and Gemini. The European community has the VLT. It is clear that finding the best international site is paramount. It is then our prime concern to make access to and participation in the greater project a National priority. Why should we build the RSST? _____________________________ We have established a number of arguments for US investment in a next-generation radio array as a step in a national plan for the advancement of our knowledge and research capabilities in astrophysics and cosmology. These arguments can be applied to any of the SKA concepts, including an EOR experiment and high-frequency centimeter-wave array in addition to an RSST. Based upon discussions within the astronomical and physics communities, the various forward-looking reports written for the guiding bodies of our national community, and presentations given in and heard at public forums, we have put together this proposal for a specific realization of a SKA: the RSST. We feel that this instrument concept best encompasses the interests and expertise of the wider scientific community that has a stake in large future infrastructure investments. Participating in the RSST project as a leading or significant partner will give our astronomers access to a premier observing resource, as well as spurring technical development in cutting edge hardware and software capabilities. Student training will be strongly supported by this project, promoting a new generation of leaders and doers. The RSST is a new kind of telescope, one that is more than just steel and silicon. In many ways, it is a telescope made of software as it is hardware. Its success will be spurred on by, and in return will foster, new developments in the technologies and techniques of information processing and storage. The RSST touches upon all the core technologies that the future health of our national research infrastructure relies upon. What will the RSST do? ______________________ The primary science driver for the RSST is a Hydrogen line survey of the Universe from z=0 to z=2.5, with the ability to detect and resolve Milky Way like galaxies with resolutions of 1 arcsec or better. As a straw-person concept, we designate this the Cosmological HI Large Deep Survey (CHILDS). The CHILD survey will allow the determination of the redshifts of the spiral galaxies in the cosmological volume probed by surveys done with the next generation of optical/infrared telescopes. The RSST will have other astronomical science products. The RSST will produce a record of the HI content of galaxies over the past several Hubble times, leading to a tremendous advancement in our understanding of galaxy evolution. It will produce extremely sensitive continuum images of the radio sky in the frequency band 0.4-1.4 GHz. It will be sensitive to transient radio sources due the synoptic survey mode of data acquisition. Although its back-end processing will be designed for imaging, there will be the capability to attach special purpose instruments to a "spigot", such as for pulsar timing and detection. See the SKA Science Book (http://www.skatelescope.org/pages/page_sciencegen.htm) for many examples of the transformational science that will be made possible with a next generation radio array such as the RSST. When could we build the RSST? _____________________________ The funding of a project such as the RSST would begin in the next decade (2010 and on). There are a number of small research and development efforts being funded this decade, mostly under the US-SKA Technical Development program, but also through independent projects such as the Allen Telescope Array (ATA) project, and the Expanded Very Large Array project. There are also individual researchers working on new technical and algorithmic approaches necessary to the success of an RSST. In our RSST concept, which is focused on mapping extragalactic neutral hydrogen to redshift z~2, there are no obvious technical show-stoppers that need a decade of development to realize. There are significant developments that will need to happen to reduce the costs of some aspects and to optimize the efficiency and sensitivity of the array. This is a do-able project. How would the RSST be built? ____________________________ This white paper is intended as a case for RSST science. However, we also note that there are technical challenges in realizing this instrument, mostly due to the large number of individual telescope or focal elements required (typically hundreds or thousands) and extremely large data rates and volumes output from such an instrument. These issues have been explored as part of the SKA project and we refer the reader to these exhaustive studies. There are a number of technologies available for the construction of an RSST. These have been put forth in the context of the SKA in a series of technology white papers: http://www.skatelescope.org/pages/page_astronom.htm For angular resolution of 1 arcsec at 400 MHz, interferometer baselines up to around 150 km are required. Baselines up to 3000 km can be accommodated by any of the sites (e.g. Australia). Leading concepts applicable to the RSST are: o an array of a large number of "small" parabolic dishes (LNSD, the US SKA concept, based on the ATA or GMRT ) o an array of a moderate number of large cylindrical elements (the Australian concept, based on Molonglo/SKAMP, now also being investigated in the US) o an array of an extremely large number of aperture tiles (the European SKA concept, based on LOFAR) These concepts are under investigation, and the design studies will come to fruition by the end of the decade. In summary, the RSST would: o cover frequency range 400-1400 MHz (z=0 to z=2.5) o have 10^6 m^2 of collecting area or more, with reasonable Tsys: the fiducial SKA Aeff/Tsys = 2 x 10^4 m^2/K o be made up of a large number of small apertures (LNSD) - if D=12m then A=100 m^2 and N=10^4 elements are needed - if D=51m then A=2000 m^2 and N=500 elements are needed currently the 12m is the favored design (or even 6m). Note that 25m VLA-style aperture sizes are not out of the question and would be a compromise, with A=500 m^2 and N=2000. o need a massive correlator - N=1000 -> 10^6 correlation pairs (real + imaginary) - 30 km/s resolution = 10^-4 spectral resolution -> ~ 10^4 channels - 2 or 4 polarizations - high time sampling >10Hz (though can integrate up to 1s later) - Ncor = 10^6 x 10^4 x 2 x 10 = 200 Gx per sec o probably have <1 arcsec resolution at 400 MHz (baselines > 130 km) o have moderately large field-of-view - for D=12m the instantaneous field-of-view is 1deg/3deg (1400/400MHz) - for D=51m the instantaneous field-of-view is 0.24deg/0.72deg (1400/400MHz) o operate in survey mode to cover large area of sky (10^4 sq.deg) often (synoptic) - FOV for 12m ~ 1 sq.deg => need to map 10^4 resolution elements - could cover in 1 day if dwell ~8s per beam o A reasonable target is to have 10^6 square meters with 1 square degree FOV. This means that on average you have 10^2 square meters on every pointing to cover 10^4 square degrees (1/4 of the sky), the equivalent of a 12m telescope. o be located in Western Australia (lowest RFI site) or possibly in South Africa o have an extremely large data rate and volume for processing. Efficient and novel processing algorithms will be required, along with massively-parallel processing and I/O computing capabilities. In addition, the RSST o will likely have a staged architecture from the elements (antenna+receiver) through the correlator to the archive and finally post-processing pipeline: array element (antenna/receiver/digitizer) | V back-end interface (fiber-optics/MUX/ethernet spigot) | V correlator / ABE (correlator/integrator) | V archive (raid farm/parallel I/O) | V post-processing pipeline (cluster/software) | V science archive (raid farm/parallel I/O /web interface) The hardware, computing and information technology development for each of these stages is considerable, and of considerable interest to the scientific and engineering community in the US. How would the RSST be operated? _______________________________ The primary use of the RSST would be to carry out a Cosmological HI Large Deep Survey (CHILDS). To reach the required sensitivity levels to detect Milky-Way like galaxies (M* = 5 x 10^9 Msun in HI) at redshifts from 1 out to 2.5, integrations on the order of 4 hours to 360 hours are required respectively for a canonical SKA (Abdalla & Rawlings 2005). The RSST could be optimized for this band with somewhat increased collecting area and system performance. The survey would likely be carried out to a shallow depth (4h integration, z<1) over a large area (10^4 square degrees) and to maximal depth (350h integration, z<2.5) of a smaller area. The total integration time per field on the sky would be built up over the course of time by revisiting each field with as rapid a cadence as possible, perhaps daily for the deep areas and correspondingly longer for the shallow areas. The RSST is a supremely powerful telescope, reaching a thermal sensitivity rms of 0.5 nJy or deeper in 1s in a 40 kHz bandwidth (30 km/s at 400 MHz). Thus even short integrations will be of astronomical interest. We envision that the RSST would be first and foremost an HI survey instrument, with at least 90% of its time devoted to the CHILDS. This is similar to how the SDSS operates. The remaining time could be used for other surveys or for general observer (GO) proposal-driven observations. By focusing on CHILDS mode operational costs can be controlled. If significant GO observations and user support are desired by the community, then extra operational support should be provided externally to the project. What are the RSST Design and Development issues? _______________________________________________ There are a number of topics involved in the design and development of an RSST concept that could significantly involve US scientists and engineers. The RSST would ultimately involve a large international collaboration on a S1G instrument over a timescale of a decade or more. Much of the cost is computing and software to handle the 1000's of interferometer elements and huge data rates and volumes, which could naturally incorporate technologies developed by the NSF and DOE funded university, laboratory and observatory based groups as the core data and science processing for RSST. The RSST should make use of existing radio arrays for prototyping and studies. These include the: o Expanded Very Large Array (EVLA), sited in New Mexico. (http://www.aoc.nrao.edu/evla/) o Long Wavelength Array (LWA), sited in New Mexico with a core at the VLA. (http://lwa.unm.edu/) o Allen Telescope Array (ATA), sited in California at Hat Creek. (http://ral.berkeley.edu/ata/) Nearer term topics that might form the basis for small projects and proposals include, but are not exclusive to: o Design and development of the CHILD Survey proposal, including the requirements on RSST capabilities. This would greatly expand on the ideas presented in this document, and would bring the CHILDS/RSST concept to the level needed for presentation to the next Decadal Panel. Wide community participation is desirably, preferably through a core of key individuals tasked with putting this together. o Design and development of the RSST array concept. This would involve taking the SKA design concepts and optimizing them for the RSST and the CHILD Survey in particular. o Design and development of an Agile Back End (ABE) correlator spectrometer for RSST. This could be based on use of commodity computing or video processors. Prototype applications could be for use on the LWA and/or ATA. o Development of high-performance data processing pipelines for calibration and imaging. This would require massively parallel I/O and processing. Though ultimately for RSST, would be prototyped for LWA and EVLA and usable for ATA. o Ionospheric calibration and reconstruction. The ionosphere will be the major "climatic" limitation on performance of LWA and RSST (and EVLA below 2 GHz) and must be dealt with. With an array size of 100 km or larger, LWA and RSST will essentially image the ionospheric volume above the array, and tomographic reconstruction can image that volume along with the sky. o Non-imaging data mining of the RSST archive, including transient (burst) detection and pulsar processing, power spectrum estimations, as well as interference (RFI) detection and mitigation. Prototyping can be done on ATA, EVLA and LWA. o Object detection and catalog construction in RSST archive. Note that simple image-based algorithms are inadequate for interferometry. This is in 3-D given the spectral dimension. Builds on work for surveys such as with SDSS and VLA. Prototype on ATA, EVLA, and LWA. o Design studies based upon simulation of HI in galaxies and large-scale structure as a function of cosmology. This includes combining RSST and LSST and JDEM (or equivalent DE survey) simulated data. o Weak lensing science from RSST data. o Other astrophysics with RSST data (AGN studies, GRB detection, SNR discovery and mapping, galactic and local group studies). Note that the International SKA and US SKA studies that have been done on these topics are excellent starting places for RSST studies, and in many cases may be "good enough". What needs to happen next? __________________________ For the RSST to happen, a more fully fleshed-out White Paper must be submitted to the Decadal Panel, which will presumably be established in 2008. The process is still being decided on, see http://www7.nationalacademies.org/bpa/astronomy_and_astrophysics_decadal_survey.html for more information. The key things that need to happen before then are: o Get together a small "blue team" to put together the White Paper. o A reasonable amount of community input needs to inform the Blue Team, including the US SKA Consortium as well as from interested astronomers (from all wavelengths). But time is short, and much of the science and technology cases have been made previously. o Some decisions on readiness to commit to technology options must be made. Can hopefully rely upon some consensus of the US and International SKA projects, as long as they are informed and backed up by solid arguments. o Some linkage to related optical/infrared projects should be explored (e.g. LSST, JDEM). This will be particularly important. o Make the case as broad as possible, so that support from all relevant agencies (NSF, DOE, NASA) can be garnered. o Reasonable costing must be done. This looks to be particularly important for the next Decadal Review (see discussion on the website). o Write the thing! Looks like we have only 1 year or so. References: ___________ "Building the Foundation for U.S. Astronomy at m/cm Wavelengths in 2010 and Beyond" Community Workshop, August 3-4, 2006, Tucson, AZ (http://www2.naic.edu/~astro/chicago2/) Recommendations and Actions (http://www.naic.edu/~astro/chicago2/recommendations_actions.pdf) The Radio, Millimeter and Submillimeter Planning Group (RMSPG) (http://www.astro.cornell.edu/~haynes/rmspg/) Final Report (http://www.astro.cornell.edu/~haynes/rmspg/docs/rmspgreport.pdf) "Connecting Quarks with the Cosmos: Eleven Science Questions for the New Century (2003)" Board on Physics and Astronomy (http://www.nap.edu/books/0309074061/html/) Dark Energy Task Force (DETF) Final Report (http://www.science.doe.gov/hep/DETF-FinalRptJune30,2006.pdf) (http://www.nsf.gov/mps/ast/aaac/dark_energy_task_force/report/detf_final_report.pdf) "Probing dark energy with baryonic oscillations and future radio surveys of neutral hydrogen" Abdalla, F. B. and Rawlings, S. (2005), MNRAS, 360, 27-40. "Galaxy evolution, cosmology and dark energy with the Square Kilometer Array" Rawlings, S. et al. (2004), New Astronomy Reviews, 48, 1013-1027 (http://www.skatelescope.org/PDF/sciencebook/5Rawlings.pdf) "Cosmology with the SKA" Blake, C. A., Abdalla, F. B., Bridle, S. L. and Rawlings, S. (2004), New Astronomy Reviews, 48, 1063-1077. (http://www.skatelescope.org/PDF/sciencebook/09Blake.pdf) "Motivation, key science projects, standards and assumptions" Carilli, C. L. and Rawlings, S. (2004), New Astronomy Reviews, 48, 979-984. (http://www.skatelescope.org/PDF/sciencebook/Carilli.preface.pdf) "The SKA: an engineering perspective" P.Hall - Springer, 2005 reprinted from Experimental astronomy, Vol 17, Nos 1-3, 2004 (http://www.skatelescope.org/pages/page_documents_EngBook.htm) The Large Synoptic Survey Telescope (LSST) (http://www.lsst.org) The Supernova / Acceleration Probe (SNAP) (http://snap.lbl.gov/) The Expanded Very Large Array (EVLA) (http://www.aoc.nrao.edu/evla/) The Long Wavelength Array (LWA) (http://lwa.unm.edu/) Allen Telescope Array (ATA) (http://ral.berkeley.edu/ata/) The Australian Square Kilometre Array Pathfinder (ASKAP) (http://www.atnf.csiro.au/projects/mira/index.html) The MWA (http://www.haystack.mit.edu/ast/arrays/mwa/) SKA South Africa and meerKAT (http://www.ska.ac.za/)