Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.naic.edu/alfa/ealfa/meeting1/talks/staveley-smith-HIPASS.pdf Дата изменения: Mon May 8 22:43:33 2006 Дата индексирования: Sun Dec 23 05:20:37 2007 Кодировка: Поисковые слова: обвмадеойс нефептощи рпфплпч

HIPASS ­ RESULTS & LESSONS
! A HIPASS result ­ Bright Galaxies and the HIMF ! The ALFA advantage ­ Speed, depth, resolution ! Things to get right ­ On-line & reduction software, obs strategies ! Things to think about ­ INTMIT, deconvolution
Lister Staveley-Smith ­ X-ALFA March 15, 2003 1


HIPASS Bright Galaxy Catalog (BGC)
Koribalski et al. (2003)

! HIPASS has

detected ~7000 galaxies with Decl.<25o ! Bright Galaxy Catalog defined to contain the 1000 HI-brightest galaxies with Decl.<0o (S>116mJy)


New HIPASS bright galaxies
Ryan-Weber et al. (2002)

·87/1000 previously uncatalogued galaxies ·138 new redshifts


BGC 2DSWML joint HI massvelocity function
Zwaan et al. (2003); astro-ph/0302440

BGC HI mass function


Comparison with previous work
(see also Steve Schneider's talk)

Lister Staveley-Smith ­ X-ALFA March 15, 2003

5


The ALFA advantage - I
! Speed
­ e.g. 7 days would complete Decl. range 25o-38o to HIPASS sensitivity!

ALFA PKS



SALFA nPKS PKS = SPKS nALFA ALFA
0.64 K .Jy -1 28 K в 21K 9 K .Jy
-1

2

0.26
14'.3 3'.6
2



2

13 7

sensitivity

#beams

beam area

Time required to survey a given region to a given sensitivity (see also D.J. Pisano's talk)
Lister Staveley-Smith ­ X-ALFA March 15, 2003 6


The ALFA advantage - II
! Sensitivity ­ HIPASS (=250 sec/pix) is complete for M* galaxies only for cz<4000 km s-1. ­ ALFA (same ) would be complete to cz<15,000 km s-1 ! ­ Deep ALFA (=few hrs) should be able to detect M* galaxies to z=0.1-0.16 (the max ALFA redshift) ­ Don't forget single-object `MX' mode (object always in a beam) which is up to 4 TIMES FASTER than for single-beams
HI at z=0.18 in a A2218 spiral with WSRT (Zwaan et al. 2001)
Lister Staveley-Smith ­ X-ALFA March 15, 2003 7


The ALFA advantage - III
! Resolution
N.B. for a filled aperture, linear resolution always the same at the max survey depth dmax (A/T )1/2.
Arecibo beam Parkes beam

HIPASS spectrum

HIPASS J0620-57 = ESO161-G001/NGC 2222/NGC 2221 (ATCA observation)
Lister Staveley-Smith ­ X-ALFA March 15, 2003 8


The ALFA advantage ­ an example
! Cosmic variance for HIPASS BGC dwarfs

#10 Mpc cells

Dwarf galaxies
Lister Staveley-Smith ­ X-ALFA March 15, 2003 9


Cosmic Variance in HIPASS BGC

Zwaan et al. (2003)
Lister Staveley-Smith ­ X-ALFA March 15, 2003 10


Things to get right...
! Correlator ­ Ideal would be 4-8 bits x 200 MHz x 8192 channels x 2 polarisations x 7 beams + INTMIT [present plan 1.5 bits x 100
MHz x 4096 channels x 2 polarisations x 7 beams]

! On-line software ­ Hide complexity
calibration ...)

(parallactification, scanning schemes,

­ Allow flexibility (allow users to experiment) ­ Schedulable or interactive observing

! Reduction software ­ As above! ­ Needs to be able to run in real-time
Lister Staveley-Smith ­ X-ALFA March 15, 2003 11


Example
! ! ! ! ! ! ! ! !

Let ALFA re-map a 64 deg2 HIPASS field Use HIPASS sensitivity (13 mJy/beam) Require Nyquist sampling of sky with all 7 beams Bandwidth 100 MHz, 7x2x4096 channels Integration time = 1 sec/beam; Total time = 2.6 hrs Drive rate = 20 deg/min [IMPOSSIBLE IN ZA] Correlator integration period = 60 ms [OUTSIDE SPECS] Minimum data volume = 36 GB Minimum data rate = 4 MB/s (=200xHIPASS)

! The Parkes LiveData processor will NOT handle this ! The JCMT/DRAO ACSIS 32-processor linux cluster might
Lister Staveley-Smith ­ X-ALFA March 15, 2003 12


Things to think about - I
! Interference mitigation (Briggs, Bell & Kesteven 2001)

t
A B
Raw multibeam spectra near 1499 MHz

A

B



Post-correlation RFI cancellation with reference horns
13

Lister Staveley-Smith ­ X-ALFA March 15, 2003


Things to think about - II
! Deconvolution

Nearby extended sources, or confused regions will need deconvolution. Complicated by a PSF which varies with time, frequency, position and beam number.
Cortes-Medellin (2002)

Lister Staveley-Smith ­ X-ALFA March 15, 2003

14


Summary
! HIMF possibilities (faint-end, cosmic variance, density dependence...) ! Large-scale, shallow ALFA surveys not

possible. ! Correlator capability a bit low for surveys and INTMIT requirements ! Data reduction bottleneck? ! Deconvolution methods ...
Lister Staveley-Smith ­ X-ALFA March 15, 2003 15