Observing strategies Max Voronkov Software Scientist ASKAP 28

Observing strategies Max Voronkov Software Scientist – ASKAP 28 th September 2010

Motivation In legalese: No student or students, or persons acting on the direction or suggestion or supervision of the student or students may try, or attempt to try or make an attempt to try to use the facility without basic understanding of the technique, except as authorized by the head of the science operations. The offence against this regulation is an offence of strict liability. The penalty is 1 proposal grade unit. In plain English: You can do a better job if you’re know what you’re doing

ATCA and CABB Practical issues for an ATCA user May be handy for other telescopes as well

Writing your proposal There a number of questions about your observations which must be answered at the proposal preparation stage

Proposal preparation - 1 • What to observe? • • When? For how long? Frequency? Array configuration? Image: Kim et al. Image: White & Duncan Image: Saripalli et al. What to observe largely depends on your science! You may want to narrow down the selection of targets to maximize the science return thinking about • How long the source can be observed with ATCA? • Shadowing? • Synthesized beam quality (bad for equatorial sources) • Overheads (e. g. bad for sources transiting through zenith) • Other issues like confusion and dynamic range

Proposal preparation - 2 • What to observe? • • When? For how long? Frequency? Array configuration? Image: Kim et al. Image: White & Duncan Image: Saripalli et al. For a variable source science may dictate when to observe! Even if variability is not important, it is often worth to consider whether the target is visible during the day or at night • Better atmosphere at night for mm-wavelength observations • Less solar (and probably man-made as well) interference at night (for low frequency observations) • Afternoon thunderstorms in summer

Proposal preparation - 3 • What to observe? • • When? For how long? Frequency? Array configuration? Image: Kim et al. Image: White & Duncan Image: Saripalli et al. Check the sensitivity calculator! http: //www. atnf. csiro. au/observers/docs/at_sens Wait for lecture on sensitivity if you want the math • Effective bandwidth and spectral resolution • uv-coverage: spread short integration across the range of hour angles (note, overheads are increased!) • Overheads (understanding comes with experience…)

Proposal preparation - 4 • What to observe? • • When? For how long? Frequency? Array configuration? Image: Kim et al. Image: White & Duncan Image: Saripalli et al. Usually the selection of frequency is driven by science! If you have a choice consider how the telescope performs at different frequencies

Proposal preparation - 5 • What to observe? • • When? For how long? Frequency? Array configuration? Image: Kim et al. Required resolution Image: White & Duncan Image: Saripalli et al.

Proposal preparation - 6 • What to observe? • • When? For how long? Frequency? Array configuration? Image: Kim et al. Image: White & Duncan Image: Saripalli et al. • How big is the object / area of interest? • Single pointing: field of view is ~ /D (primary beam) • Mosaicing: stitching several pointings together Image (rotated Cen A): Cornwell & Feain • uv-coverage per pointing • overheads • single dish data? • Tasks mosgen and atmos in MIRIAD will help to setup

Proposal preparation - 7 • What to observe? • • When? For how long? Frequency? Array configuration? Image: Kim et al. Image: White & Duncan • How big is the object / area of interest? • Single pointing: field of view is ~ /D (primary beam) • Mosaicing: stitching several pointings together • Spectral line or continuum? Zoom Image: Saripalli et al.

CABB - range of modes • Two independent 2 GHz wide spectral windows (or IFs as they’re often called) • Plus optional zoom windows (now up to 4 zoom windows per 2 GHz window) • Coarse resolution of the wide (2 GHz) window = bandwidth of 1 zoom window • Optional stitching of zoom windows is possible

Actual observations Consider the simplest case: one band from 4. 5 to 6. 5 GHz • Go to 1934 -638 (flux calibrator) (it is strong enough to do all the required calibration) • Initial array calibration (delays, phases and amplitudes) • Close the file (junk data) • Track 1934 -638 for bandpass and flux calibration, e. g. 10 min Then, start the loop with program source(s): Nearby calibrator (<10 o) for e. g. 2 min Target for e. g. 20 min ……… If you worry about bandpass variations, observe bandpass calibrator several times (e. g. at regular intervals). Note, phase calibrator may be good enough!

Initial calibration (dcal, pcal, acal) vis display Why bother? Delay = phase slope Frequency average leads to decorrelation Phase calibration (“corr pcal a”) Delay calibration (“corr dcal a”) Amplitude calibration (“corr acal a”)

Variations - 1 • A bit lower frequencies (i. e. near 1. 4 GHz) • • • The second 2 GHz band (IF) is redundant Only part of the band has signal (at this stage) Many interference spikes Need to choose the channel range for initial calibration (via “corr tvch”) Calibrators are often confused (hard to find isolated point sources) • A bit higher frequencues (i. e simultaneous 4. 5 -6. 5 GHz and 8. 0 -10. 0 GHz) • 1934 -638 may be too week for bandpass (about 2 Jy at 9 GHz, which is not too bad, but is getting low) • Use something else for bandpass and initial setup (i. e. mm-wavelength bandpass calibrators 1253 -055, 1921 -293 and 0537 -441) • Have to give a flux estimate with the “corr acal” command (doesn’t need to be very accurate as the flux scale is corrected during the data reduction, but it is good to have the Tsys scale roughly right) • Still observe flux calibrator 1934 -638 at least once (to fix the flux scale)

mm-wavelength observations • 1934 -638 is definitely too weak for bandpass at mm-wavelength • Use mm-wavelength bandpass calibrators 1253 -055, 1921 -293 and 0537441 for bandpass/initial setup • Have to give a flux estimate with the “corr acal” command at 12 mm and 7 mm (3 mm is special, no “corr acal” is needed, but need a paddle scan) • 1934 -638 is still a flux calibrator at 12 mm and 7 mm. Observe a planet at 3 mm (normally Uranus). • There is an additional step to adjust power (mm-attenuators) • • You may need to play with them during setup (e. g. “set mm ca 01 3 3”) Setting is per polarisation (applies to both frequencies) The goal is to move CABB attenuators away from the edge (0 or 15) May be tricky if two 2 -GHz IFs are well separated • Need a pointing scan as the global pointing model is not good enough • Observations of continuum calibrator with half-beam offsets • Close to the target (<15 o). Don’t do pointing on a planet. • Roughly once an hour or if moving to a different part of the sky. • May need to adjust focus

Pointing scan • Pointing scan offsets dishes around expected position • Options to choose suitable initial estimate: update, offpnt, refpnt • caobs options: point_antenna, point_ifflag and point_pattern

Paddle scan • Use paddle to measure the system temperature and correct for the atmosphere at 3 mm • • • Kate’s talk for more details…. Paddle replaces the amplitude calibration (“corr acal” during setup) Every 15 -20 min or if moving to a different part of the sky or change frequency Do it near phase calibrator / target source Also need paddle scans before flux and bandpass calibrator scans

Weather Calibrator scans No data is often better than bad data! Atmosphere attenuates astronomical signal and adds noise! In addition, the effect is not constant in time

Typical 12 mm or 7 mm observation I often like to add setup sources (bandpass and flux calibratior + associated pointing scans) to the same schedule and delete them just before I start the loop • Can benefit from the “Master/Slave” feature • Can just copy the scans and ensure exactly the same parameters are used for your setup scans • Adding scans on calibrators is very easy (the scheduler knows everything about these sources, you just type the name)

Typical 3 mm observation It may be handy to create a separate schedule for setup sources (but same approach as before can also be used)

Special considerations for line observers We put sky frequencies into ATCA schedules (in other words we work in the topocentric reference frame) Science is usually done with the spectra in either LSR or barycentric reference frame LSR = local standard of rest (motion w. r. t. nearby stars) Barycentric is w. r. t. barycentre of the Solar system Conversion is position- and time-dependent May be tricky for narrow-band observations of multiple sources!

Polarisation observations CABB always measures XX, XY, YX and YY. But we need to know the phase offset between X and Y as well as the leakage between them to convert the measurement into something scientifically useful. Need unpolarised/weakly polarised source observed at a number of hour angles to estimate instrumental polarisation 1934 -638 has <0. 2% polarisation, phase calibrator may be good enough Rely on absolute XY phase measurement using noise diode, otherwise we would need an additional bright and strongly polarised calibrator Need good bandpass calibration. Additional challenges: direction-dependent polarisation (primary beam), variation of instrumental polarisation within 2 GHz band

Further information • Check the users guide • http: //www. narrabri. atnf. csiro. au/observing/users_guide/html/atug. html • Check the ATCA forum / post questions • http: //atcaforum. freeforums. org/ • Come to observe and learn from the experience • Talk to your Duty Astronomer and experienced observers • Check current issues • http: //www. narrabri. atnf. csiro. au/observing/Current. Issues. html • Check CABB pages • http: //www. narrabri. atnf. csiro. au/observing/CABB. html

Australia Telescope National Facility Max Voronkov Software Scientist (ASKAP) Phone: 02 9372 4427 Email: maxim. voronkov@csiro. au Web: http: //www. narrabri. atnf. csiro. au/~vor 010 Thank you Contact Us Phone: 1300 363 400 or +61 3 9545 2176 Email: enquiries@csiro. au Web: www. csiro. au