Exploring the HI Universe with ASKAP
EExploring the HI Universe with ASKAP
Martin Meyer ∗ International Centre for Radio Astronomy ResearchM468, The University of Western Australia35 Stirling Highway, Crawley 6009, AustraliaE-mail: [email protected]
The DINGO team ∼ mmeyer/dingo The survey speed of A
SKAP makes it a prime instrument with which to survey the H I universe,enabling it to carry out both wide surveys of the entire sky, as well as deep surveys covering cos-mologically representative volumes. Here, the use of A SKAP to study deep H I fields is discussedas proposed by the Deep Investigation of Neutral Gas Origins (D INGO ) survey. This A
SKAP sci-ence survey project anticipates observing in excess of 10 sources out to redshift z ∼ .
4. Keyscience goals include: Ω HI and its evolution, the cosmic web as traced by distributions such as theH I mass function and the 2pt correlation function, and the formation and evolution of galaxies.Science returns are maximised by targeting the G AMA survey regions, enabling the H I content ofgalaxies to be studied and understood in full context with all the major galactic constituents overthe past 4 Gyr. Panoramic Radio Astronomy: Wide-field 1-2 GHz research on galaxy evolution - PRA2009June 02 - 05 2009Groningen, the Netherlands ∗ Speaker. c (cid:13) Copyright owned by the author(s) under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike Licence. http://pos.sissa.it/ a r X i v : . [ a s t r o - ph . C O ] D ec INGO
Martin Meyer
1. Introduction
Understanding how galaxies form and evolve is one of the key astrophysical problems of the 21 st century. Recent observations of galaxies, supernovae, and the Cosmic Microwave Backgroundhave refined measurements of key cosmological parameters and simulations have accurately mod-elled how dark matter halos collapse from the miniscule density perturbations carried over fromthe inflationary era. But understanding how baryons collapse into these potential wells and ulti-mately form stars, how black holes are formed, how the baryons interact with the dark matter, whatthe balance is between accretion and outflow, heating and cooling, are problems which require adeep understanding of physics on an immense range of spatial scales and particle densities. Theseproblems remain largely unsolved.One of the keys to a deeper understanding is to observe and model the gaseous component ofthe Universe. Baryons are believed to have existed entirely in gaseous form before the epoch ofgalaxy formation, and it is by the continual collapse of this gas into the filamentary structures ofthe cosmic web, its accretion into the deepest potential wells, the creation of cool gaseous disksand their ultimate collapse into dense molecular clouds, that galaxies are able to form stars. Theon-going feedback between gas and stars, and the governing role of the host dark matter potentials,continue to determine the evolution of galaxies today. To date, our ability to observe the gaseouscontent of the universe, and hence to explain the manner in which these processes have occurred, isextremely limited. Observations have largely been restricted to small samples of gas in emission inthe local universe, or to sparse samples of absorption line systems at higher redshifts, and the highlevel of uncertainty inherent in interpreting such observations. D INGO and A SKAP
The Australian S KA Pathfinder telescope (A
SKAP ) [1] has been designed as a survey instrumentand is particularly well-suited to the search for 21cm neutral hydrogen emission. Its large field ofview (30 deg ), in combination with its collecting area (4072 m ) and system temperature (50K)give it a fast survey speed, enabling the telescope to go both wide and deep in a comparativelyshort period of time. Its wide 300 MHz instantaneous bandwidth similarly facilitates surveys ofcosmologically representative volumes in H I .Where W ALLABY will perform a shallow survey of the southern sky, D
INGO aims to providea legacy deep dataset for H I emission out to z ∼ .
4. To enable multiwavelength studies, D
INGO will target the G
AMA survey areas [2]. G
AMA is composed of several imaging and spectroscopicsurveys using the world’s pre-eminent facilities, spanning all wavelengths from the ultraviolet tothe far infrared (G
ALEX , V ST , V ISTA , U
KIRT , W
ISE , Herschel-A
TLAS ), in addition to 250k AA Ω spectra in key overlap regions. A tiered H I survey strategy is proposed covering regions 150 deg (500 hours per pointing, 0 . < z < .
29) and 60 deg (2500 hours per pointing, 0 . < z < .
43) tomeet the science goals described in the following section.D
INGO is a major S KA pathfinder experiment and will provide an ideal dataset for S KA plan-ning. This will be acheived scientifically through its study of the evolving H I universe, and tech-nologically through the use of A SKAP ’s phased-array feeds for extended integrations.2
INGO
Martin Meyer
Figure 1:
Simulated spatial distribution of galaxies in a 2500 hr, 30 deg A SKAP pointing. Our proposedultradeep low-z data includes two fields 0 . < z < .
43 (2500 hr integrations) while the deep includes fivefields 0 . < z < .
29 (500 hr integrations) to provide adequate cosmic volumes and galaxy numbers fordependency analysis.
3. Key Science Goals
The core D
INGO science themes and goals are: • Evolution of the cosmic H I density: On the largest scales, evolution of the gaseous uni-verse is traced by measuring changes in the cosmic gas-mass density with time. The H I Parkes All-Sky Survey (H
IPASS ) [3] and other large blind H I surveys have provided accuratemeasurements of the local H I cosmic mass density [4]. However, beyond z = I detection has limited our ability to detect H I in emission in individual galaxies,necessitating the use of stacking techniques [5] and a hetergeneous mix of methods at thehighest redshifts [6, 7] with significant uncertainty in both measurement and interpretation.D INGO aims to measure Ω HI directly through H I emission with H IPASS -like errors out to z ∼ . • The cosmic web: As well as changes in the H I cosmic mass density, it is also important tounderstand the way in which the distribution of H I gas has changed with time. This has var-ied significantly over the history of the universe, evolving from a near uniform distributionwith only small density fluctuations at the start of the dark ages, to its current highly clus-tered state, mostly restricted to galactic halos where its density is high enough to self-shield.D INGO will examine the most recent changes in the H I mass distribution through evolution-ary measurements of the H I mass function [4], the 2pt correlation function [8], and HODmodelling [9]. The simulated spatial distribution of galaxies in a single D INGO ultradeepfield is shown in Figure 1. 3
INGO
Martin Meyer • Galaxy evolution: Large, comprehensive datasets with galaxy-by-galaxy data will ultimatelybe required to fully understand the baryonic processes in galaxies. G
AMA is the best multi-wavelength dataset in existence over the areas and depths relevant to D
INGO , and by targetingthese survey regions, D
INGO will enable an understanding of the evolution of H I in thecontext of all the other major galactic constituents over the past 4 Gyr. Our models based oncurrent A SKAP specifications indicate that D
INGO will be sensitive to M ∗ HI galaxies over theentire redshift range probed.
4. Design Study D INGO is now in the design study phase of the A
SKAP survey science process. During this time, anumber of scientific and technical aspects necessary to making D
INGO a success will be examined.
Survey Design
A core task of the design study will be to determine the location of the D
INGO fields within theG
AMA survey regions. We will additionally examine other survey parameters during the designstudy to ensure that the final D
INGO configuration will best meet the aims outlined in the sciencecase using simulations, observations and analytic methods. Additional studies will be made of keyissues such as cosmic variance and the survey volumes required for environmental analysis.
Simulated Catalogues and Mock Skies
Simulated H I datasets provide a central means for evaluating the likely science outcomes forD INGO . To date, our simulations have been a combination of non-evolving semi-empirical meth-ods taking into account the locally measured cosmic mass density and the H I mass function, alongwith basic semi-analytic methods for the spatial distribution of galaxies. During the design study,we plan to advance on these through the use of fully semi-analytic methods such as the recent workof Obreschkow et al. [10]. Given the current relatively high mass limit of semi-analytic models, itmay remain necessary to combine semi-analytic and semi-empirical approaches to deliver simula-tions covering the full mass spectrum, although we will seek to advance work on this issue as faras possible. These catalogues will be developed in parallel to those proposed for W ALLABY andwe will investigate a number of simulation issues: • an analysis of semi-empirical and semi-analytic methods to yield simulations consistent withknown observational constraints on the cosmic H I density, the H I mass function, the cluster-ing properties of H I rich galaxies, and correlations between H I properties and environment • simulations that cover a range of plausible evolutionary scenarios • mock science analyses based on these simulated catalogues to examine our sensitivity, antic-ipated errors, and discriminatory power for separating various evolutionary models • assistance in the development of mock skies for A SKAP pipeline testing4
INGO
Martin Meyer
Radio Frequnecy Interference and Cube Combination
The deep observations of D
INGO make RFI an issue that will require special attention. The stackingof multiple observations will mean that RFI features may only become apparent some time afterobservations have begun. A methodology needs to be developed to deal with this given the limitedcurrent storage capability of visibility data.
Source Finding and Parametrization
Multiple approaches to source detection need to be examined and tested. It is likely that smoothingwith a matched filter or wavelet filtering may improve the detectability of point-like sources, whichwill be the vast majority of our likely detections. The processing overheads for these approacheswill need to be evaluated, as the wavelet filtering in particular can be computationally intensive.One of the primary tasks will be to test the completeness and reliability of the detections made withthe source-finder. Due to the automated nature of the processing, the source-finder needs to be ascomplete (ie. maximise the number of real sources found) and reliable (ie. minimise the number ofnon-real sources identified) as possible, and we will develop tools to measure both these quantities.D
INGO will make use of, and contribute to, the standard end-to-end simulation/imaging/analysispipeline under development by the A
SKAP computing group, which will enable us to incorpo-rate different simulations and observational strategies into our testing, and to provide uniform testresults to aid the analysis.
Early Science Commissioning
Critical to early A
SKAP commissioning are tests of: receiver performance and stability; beam per-formance and stability; calibration accuracy; techniques to ensure spatial and spectral invariance ofcalibration, including across beams; quality of sky models; dynamic range (continuum and spectralline); spectral bandpass stability (as a function of time, beam, position and polarization angle);mosaicing strategies for full beam combination; measurement of sidelobe response, including atequatorial latitudes; and the characterization of map noise as a function of angle and integrationtime. This last issue is particularly critical given the deep nature of this survey. BETA will be usedas a valuable testbed for assessing these issues.
Ancillary Datasets
Multi-wavelength data is vital for meeting the D
INGO science goals. We will monitor the devel-opment of the G
AMA datasets to ensure that the final D
INGO field locations are best positionedto take advantage of the data products that are ultimately available, and to fill any gaps that arisewhere feasible and necessary.
SKA Pathfinder Coordination
There are a large number of major radio astronomy facilities currently undergoing simultaneousdevelopment, including APERTIF, the eVLA, ATA, MeerKAT, and of course A
SKAP itself. TheA
SKAP legacy surveys will not be operating in isolation, and numerous major programs can beexpected on the other pathfinder facilities. In developing an H I survey strategy, it is important tobest target legacy surveys in ways that exploit the advantages of each facility and ideally operate in5 INGO
Martin Meyer a complimentary rather than overlapping manner. MeerKAT will also be an ideal instrument withwhich to survey distant H I , and in combination with A SKAP these facilities have the opportunity todevelop a layered survey strategy that will be able to provide a comprehensive view of the universe,starting with W
ALLABY over the largest areas ∼ × deg , D INGO over areas 60-150 deg , andMeerKAT surveying areas ∼
10 deg and less. The exact parameters and configurations of thesefuture facilities continue to evolve, as do the projects their user communities seek to carry out, andas such it is important to maintain collaborative links to ensure the overall science outcomes aremaximised. References [1] I. Feain, S. Johnston, and R. Braun,
The Australian Square Kilometre Array Pathfinder(ASKAP) an SKA pre-cursor , 2010, this proceedings .[2] S. Driver, P. Norberg, I. Baldry, S. Bamford, A. Hopkins, J. Liske, J. Loveday, andJ. Peacock,
GAMA: towards a physical understanding of galaxy formation , 2010,
Astronomyand Geophysics , , 5.12.[3] M. J. Meyer et. al. , The HIPASS Catalogue - I. Data Presentation , 2004,
Monthly Notices ofthe Royal Astronomical Society , , 1195.[4] M. A. Zwaan, M. J. Meyer, L. Staveley-Smith, and R. L. Webster, The HIPASS Catalogue: Ω HI and environmental effects on the HI mass function of galaxies , 2005, Monthly Notices ofthe Royal Astronomical Society: Letters , , L30.[5] P. Lah et. al. , The HI content of star-forming galaxies at z = 0.24 , 2007,
Monthly Notices ofthe Royal Astronomical Society , , 1357.[6] S. M. Rao, D. A. Turnshek, and D. B. Nestor, Damped Ly α systems at z<1.65: Theexpanded sloan digital sky survey hubble space telescope sample , 2006, The AstrophysicalJournal , , 610.[7] J. X. Prochaska and A. M. Wolfe, On the (non)evolution of HI gas in galaxies over cosmictime , 2009,
The Astrophysical Journal , , 1543.[8] M. J. Meyer, M. A. Zwaan, R. L. Webster, M. J. I. Brown, and L. Staveley-Smith, The weakclustering of gas-rich galaxies , 2007,
The Astrophysical Journal , , 702.[9] S. Wyithe, M. J. I. Brown, M. A. Zwaan, and M. J. Meyer, The Halo OccupationDistribution of HI Galaxies , 2010, astro-ph/0908.2854 .[10] D. Obreschkow, D. Croton, G. DeLucia, S. Khochfar, and S. Rawlings,
Simulation of thecosmic evolution of atomic and molecular hydrogen in galaxies , 2009,
The AstrophysicalJournal ,698