ALFA Pulsar Surveys Jim Cordes, Cornell University Arecibo 16 March 2003 Pulsar Consortium meeting 1-2 Nov 2002 Preliminary survey plans Hardware needs Organization of the consortium & working groups Synergies with other science goals (EALFA, GALFA) Data management (how to serve 1 Pbyte of data?, long-term archiving)

Targeted Classes of Pulsars Young, canonical pulsars (Galactic plane) Recycled pulsars (MSPs) (out of plane) High-velocity pulsars NS-NS and NS-BH binaries Pulsars ‘beyond the death line’ (radio magnetars?) Precessing pulsars Globular cluster MSPs X-and-  -ray selected pulsars Transient sources (e.g. giant pulses)

Why more pulsars? Extreme Pulsars: P 5 sec P orb G V > 1000 km s -1 Population & Stellar Evolution Issues Using pulsars to probe the ISM (gas & magnetic field) The high-energy connection (e.g. GLAST) Physics payoff (GR, LIGO, GRBs…) Serendipity (strange stars, transient sources)

t: 10 7 : 10 3 Search processing  High Performance Computing + well-organized data management 2002: single processor  200 x real time >2004: a cluster of Beowulf clusters can keep up with real time at observing duty cycle

Pulse broadening from multipath

D max vs. Flux Density Threshold Luminosity limited Dispersion limited Scattering limited ALFA

Implications: Optimal integration time: stay close to the luminosity-limited regime Fast-dump spectrometers: need sufficient number of channels so that search is not DM limited Better to cover more solid angle than integrating longer on a given direction (as long as all solid angles contain pulsars)

ALFA Pulsar Surveys I.Galactic plane |b| < b max ~ 3 to 5 deg II.Intermediate latitudes (b max  |b|  15 to 25 deg) III.Deep surveys toward specific objects - high-energy selected targets (multibeam for RFI) - extended targets (clusters, HII complexes, spiral-arm tangents) IV.Extragalactic targets - giant pulses from M33 (~24 ALFA pointings) V.Piggyback pulsar/transients survey on high b HI survey? (multiple passes) VI.Other

Nominal Parameters of Galactic Plane Survey 300 MHz bandwidth 1024 channels 64  s dump time polarizations summed ~4 bits/sample 7 beams 300 s dwell time 400 TB in 2000 hr  56 MB/s 30< l < 80 deg |b| < 5 deg 3 50%

Comparison of AO, GBT & Parkes (S min1 held fixed) Site GHz S sys Jy  MHz N ch tsts T int s S min1  Jy FWHM arcmin dT/d  hr/deg 2 AO * /N b GBT /N b PMB /N b * S sys = 3.6 Jy for Pix > 0 2.8Jy for Pix=0 ~ 2.3 Jy for new LBW

Comparison of AO, GBT & Parkes S min1 (AO) << S min1 (Parkes) S sys  N ch T S min1 dt/d  (Jy) (MHz) (s) (  Jy) (hr/deg 2 ) AO /N b =4.2 GBT /N b Parkes (N b =13)

Surveys with Parkes, Arecibo & GBT. Simulated & actual Yield ~ 1000 pulsars.

Spectrometer Requirements 300 MHz bandwidth (full feed) <0.3 MHz channels FPGA Correlator or FPGA-FFT or Polyphase filter approach Fast dump capability Polarization summing mode Needs rapid decision (this month)

II. Intermediate Latitude Survey Millisecond pulsars (z scale height ~ 0.5 kpc) High-velocity pulsars (50% escape) (scale height =  ) NS-NS binaries (typical z ~ 5 kpc) NS-BH binaries (typical z ~ few kpc ?) Search for: ~ 1500 hours (piggyback, filler time?)

Issues for Optimizing Surveys RFI management Characterization, test obs & algorithms, multibeam schemes (ALFA + other?) Diffractive ISS  multiple passes favored for low DM -t weighting for intermediate DM no action for high DM Refractive ISS  multiple passes for low to intermediate DM Nulling  multiple passes “Search” vs. “confirmation” Historically two different phases PMB: candidate density  T confirm ~T search  do two “searches” = two passes on sky

What Next? New survey simulations Population issues (PMB), NE2001 Optimize number of detections vs l,b,t,etc Design at-the-telescope survey modes Beam interlace, hour angles, feed rotation RFI studies, pilot observations, simulations Search code development (~TEMPO, not AIPS++) Data management plan Plan survey follow-up (timing, multi- )

Pulsar Consortium Working Groups Surveys (J. Cordes) Data acquisition (I. Stairs) Post processing (D. Lorimer) Data Management (S. Ransom) Follow-up observations (B. Gaensler)

Preliminary Protocols Consortium membership : –open policy early on, by application later –protection of student projects Data access: –open to all members during proprietary period –by application from nonmembers (during proprietary period) –uniform, baseline processing for legacy goal –encourage innovative new approaches Authorship: –rotating lead, equitable –all consortium members –opt out by inactive members (honor system) Follow up observations: similar to Authorship Discovery of exotica: full consortium involvement

Data Management Raw data –Local processing (inc. quicklook) –Processing at Consortium member institutions –Short and long-term archiving (disk/tape) –Central mainland location with high-bw pipe? –Database catalog system –Web based data selection Intermediate Data products –candidate lists –RFI identification –diagnostic plots Final products (catalogs, pulse profiles, timing models) Implied Linkage to the National Virtual Observatory as appropriate

Pilot Database Storage (Cornell Theory Center) Database information: Microsoft SQL Server Hosted at the Cornell Theory Center Stores both raw data and heavily processed data For the raw data simple queries will select chunks to serve out for users The processed data can be searched and analyzed with complex queries Database will be tuned to perform better for common/expected queries

Boundary Conditions etc. ALFA surveys can be viewed as part of a long-term, grander effort (“Full Galactic Census”) (LOFAR, SKA, ) ALFA surveys usher in a new NAIC mode of operation (not business as usual) RFI mitigation required and provides general purpose tools Data & data products = long term resources  data management policy & resources The scientific pie is large enough for shared glory but … A focused, concerted, committed effort is needed for (a) the best surveys (b) legacy results Exploit telescope time fully (transients, piggybacking)