Equatorial Imaging with e-MERLIN Including the Chilbolton Antenna

0.48 MB·English

by  Ian Heywood

Cosmology, Galaxy Formation and Astroparticle Physics
on the pathway to the SKA
Klo¨ckner, H.-R., Rawlings, S., Jarvis, M. & Taylor, A. (eds.)
April 10th-12th 2006, Oxford, United Kingdom
Equatorial Imaging with e-MERLIN Including the Chilbolton
Antenna
Ian Heywood, Hans-Rainer Klo¨ckner and Steve Rawlings
8
Oxford Astrophysics, Denys-Wilkinson Building, Keble Road, Oxford, OX13RH
0
0
Abstract.Wediscusstheequatorialimagingbenefitsthatarisefromtheadditionofthe25-metredishatChilbolton
2
tothee-MERLINarray.Itsinclusionconsiderablyenhancesthecapabilitiesofe-MERLINonandbelowtheequator.

n This will become particularly important in the era of ALMA and other upcoming southern hemisphere facilities.
a
Wepresent simulated observations of point sources in the equatorial region of thesky which is thetarget area for
J
many existing sky surveys. We find that the additional baselines created by the inclusion of the Chilbolton dish
4
favourably adjust thebeam shape of e-MERLINtoa more compact and circular shape, with significantly reduced
1
sidelobestructure.Puttingasidethebenefitsofincreasedcollectingarea,themodifiedbeamshapehasimplications

for more rapidly reaching a given completeness limit for equatorial surveys.
]
h
p 1. Introduction
-
o
The facilities at Chilbolton Observatory in Hampshire,
r
t UK, include a fully steerable 25-metre parabolic an-
s
a tenna which is mainly used for meteorological Doppler-
[
polarisation radar. The use of this antenna in the

1 MERLINarray(Thomasson,1986)hasbeendiscussedfor
v many years, and with the provision of a fast fibre link
7 it would become a prime candidate for inclusion in the
3 e-MERLIN1 array.
0
2 Figure 1 is a version of the diagram presented on the
. online MERLIN user guide2 which has been modified in
1
order to show the location of the Chilbolton antenna in
0
8 addition to the locations of the existing MERLIN sta-
0 tions. Antenna numbers for the existing MERLIN sta-
v: tions have also been added. As can be seen, inclusion
i of the Chilbolton antenna will boost the number of long
X
and intermediate-length baselines, as well as extending
r the north-south span of the array, facilitating the sam-
a Fig.1.MapshowingthecurrentMERLINarray,withthe
pling of a much greater range of spatial frequencies. This
locationoftheChilboltonObservatoryhighlightedingrey.
is particularly crucial for ‘snapshot’ mode observations,
and large-scale survey programmes where on-source time
may necessarily be rather brief.
In this article we demonstrate the improved uv-plane temperatures in the range20 - 33K. The 76-metreLovell
sampling of an e-MERLIN+Chilbolton array for both telescope is not included in the simulated array.
snapshot and full synthesis equatorial observations. We
present simulations demonstrating the response of both
2. Increased uv-plane coverage
arraystoanunpolarizedpointsource,anddiscusstheim-
plicationsoftheinclusionoftheChilboltonantennawhen Figure 2 shows the uv-plane coverage for the e-MERLIN
carrying out equatorial surveys. arrayplustheChilboltonantennafora5-minutesnapshot
All observations are simulated using AIPS with stan- observation(left panel)anda12-hourtrack(rightpanel).
dard imaging procedures. All simulations are at L-band, Theadditionalvisibilitiesarisingfromtheinclusionofthe
using100frequencychannelsof5MHz eachto simulatea Chilbolton antenna are shown in light grey.
1.3 - 1.8 GHz contiguous band, assuming a spectrally flat Eachofthesesimulationsisanobservationofanequa-
source. Each antenna has an efficiency of 0.8 and system torial source (with a Right Ascension and Declination of
zero). The snapshot observation has an on-source time of
1 http://www.merlin.ac.uk/e-merlin/ 5 minutes at an hour angle of zero and the 12-hour track
2 http://www.merlin.ac.uk/user guide/OnlineMUG/ has an hour angle range of -6 to +6. The latter scenario
Ian Heywood, Hans-RainerKl¨ockner, SteveRawlings: Equatorial Imaging with e-MERLIN + Chilbolton
gives the best possible uv-coverage for a source at this Table 2. Increase in mapping speed due to the inclu-
position due to the opacity of the earth. The integration sion of the Chilbolton antenna, expressedas a percentage
time assumed in both cases is 60 seconds. of the observing time required for full synthesis observa-
The key feature to note in both of these plots is that tionsusingtheregulare-MERLINarray.Thisispresented
the additional Chilbolton baselines greatly reduce the for sources with major axes of 0.6 and 0.3 arcseconds,
east-west bias of the array and increase the number of aligned both parallel and orthogonalto the major axis of
long and intermediate-length baselines. The image-plane the beam.
implicationsofthismodifieduv-planecoveragearedemon-
Simulation Factor
strated in Section 3.
0.6 arcsec, aligned 85%
0.6 arcsec, orthogonal 68%
3. Imaging considerations 0.3 arcsec, aligned 64%
0.3 arcsec, orthogonal 46%
Simulated, zero-noise dirty images of a 5 minute snap-
shot observation of an equatorial 1-mJy point source are
presented in Fig. 3. These images correspond to a 6- Gaussian-like radio source is observed its brightness dis-
antenna e-MERLIN array without (left) and with (right) tribution is essentially convolved with the central lobe of
the Chilbolton antenna, observing with the 1.3-1.8 GHz the beam. This convolution will ‘smear’ the flux density
contiguous band. Since these are dirty images and, since out over an area depending on the size of the source and
theyareobservationsofapointsource,theyrepresentthe the size of the beam, the result being another Gaussian3.
point spread function of the array. A Gaussian fitted to Given that a radio source will generally have a fixed
the half power level of the central lobe is plotted in the brightnessoverthe courseofanobservation,the observed
lower left of each panel. The dimensions of these ellipses peakbrightness(i.e.the‘height’oftheresultingGaussian)
in arcseconds are 0.42 0.10 arcsec2 (position angle = will be dependent on the size of the central lobe. Using a
20.4o, without Chilbolto×n) and 0.24 0.10 arcsec2 (posi- more compact beam, as in the case where the Chilbolton
tion angle = 33.3o, with Chilbolton).× antennaisincluded,reducesthe broadeningandincreases
The benefits of the additional baselines to the the peak brightness, allowing a flux limit to be reached
Chilbolton station are immediately apparent. The domi- faster in certain cases.
nant,nearverticalstructureandstrengthofthe sidelobes MERLIN observations of the Hubble deep and flank-
inthebeampatternaresignificantlyreducedwiththead- ing fields have shown that a typical micro-Jansky radio
dition of the Chilbolton antenna. The central lobe of the source has a characteristic size of 0.6 arcsec (Muxlow
∼
beam is also much closer to circularity. et al., 2005). The major axes of the snapshot beams are
Figure 4 shows cleaned, simulated snapshot images 0.4 arcseconds (without Chilbolton) and 0.2 arcsec-
∼ ∼
without (left) and with (right) Chilbolton, with Gaussian onds (with Chilbolton).
noise added, the level of which is calculated by the AIPS With the above considerations in mind, we present
task UVCON according to the antenna and observation quantitative estimates of the increased mapping speed
parameters. These simulations obviously assume perfect provided by the inclusion of the Chilbolton antenna in
phase and amplitude calibration which is never the case Table 2. These values have been calculated by convolv-
in practice. Similarly, Fig. 5 shows cleaned images corre- ing two-dimensional Gaussians, measuring the difference
sponding to a 12-hour track. The contour levels are ad- in the peak value for the two arrays, and assuming the
justedfortheseimagesinordertodisplaythe µJy-level sensitivity has an inverse dependence on the square root
background noise. Note that phase and ampli∼tude errors oftheon-sourcetime.Twomodelsconsistentwiththefull-
wouldyieldadditionalsymmetricandanti-symmetric(rel- synthesis beam shapes with and without Chilbolton, and
ative to the peak) noise patterns, similar to those seen in two source models with major axes of 0.6 and 0.3 arcsec
Fig. 3. (withminoraxeshalfthisvalue)areused.Thecalculation
Relevantparametersforthesefourimagesarelistedin is performedwith bothparallelandorthogonalalignment
Table 1. The values and uncertainties in the peak fluxes ofthemajoraxesofthesourceandbeam,representingthe
and integrated flux densities are determined by fitting bestandworstcaserespectivelyforarandomdistribution
Gaussians to the centre of the image using the AIPS task of source alignments.
IMFIT. The backgroundRMS values in the image are re-
turned by isolating the background region using TVWIN 4. Conclusions
and executing the AIPS verb IMSTAT.
NotethatinthecaseofFig.4thereisnoimprovement ThegeographicallocationoftheChilboltonantennaintro-
at all in the RMS achieved despite the increased collect- ducesbaselineswhichcomplementtheexistinge-MERLIN
ing area. The decrease in RMS expected because of the 3 TheconvolutionofanytwoGaussiansisanotherGaussian,
10% increase in collecting area is masked by sources of
since (i) the product of any two co-centred Gaussians is a
∼
systematic error. Gaussian, (ii) the convolution of any two functions is equal
Assuming that the sidelobe structure can be success- totheFouriertransformoftheirproduct,and(iii)theFourier
fullyremovedbywhatevermeans,thenwhenanextended transform of a Gaussian is another Gaussian.
Ian Heywood, Hans-RainerKl¨ockner, SteveRawlings: Equatorial Imaging with e-MERLIN + Chilbolton
Fig.2.Theuv-planecoveragesforasnapshot(left)and12-hour(right)observationofaskypatchatRA=0andDec
= 0 with an hour angle range of -6 to +6. The additional visibilities due to the inclusion of the Chilbolton antenna
are shown in grey.
Fig.3.Simulated,zero-noisedirtyimagesofa5-minute snapshotobservationofa1mJy pointsourceatRA=0,Dec
= 0 and hour angle = 0, with e-MERLIN without (left) and with (right) the Chilbolton antenna. These images are
equivalentto the dirty beams of the two arrays.Contourlevels are -30%,-20%,-10%,10%,20%,... ,90%,100%times
the peak in the map.
Table 1. Cleaned image parameters for Fig. 4 and Fig. 5, recalling that the input model is a 1-mJy
pointsource.B andB arethemajorandminoraxesrespectivelyoftheellipsefittedtothecentral
maj min
lobe of the beam at the half-power level.BPA is the position angle of this ellipse. eM in the simulation
column indicates use of solely the six e-MERLIN antennas whereas eM +C means that the Chilbolton
antenna is added. The RMS increase due to the increases collecting area is apparent when comparing
the values corresponding to the 12-hour tracks.
Simulation Peak flux Integrated flux Background RMS Bmaj Bmin BPA
(mJy/beam) density(mJy) (mJy/beam) (arcsec) (arcsec) (deg)
Snapshot (eM) 0.903 ± 0.044 1.081 ± 0.085 0.056 0.422 0.104 20.4
Snapshot (eM+C) 0.936 ± 0.063 1.024 ± 0.115 0.058 0.241 0.108 33.3
12 hour(eM) 1.006 ± 0.007 1.025 ± 0.012 0.009 0.297 0.138 22.5
12 hour(eM+C) 1.013 ± 0.007 1.009 ± 0.011 0.007 0.174 0.122 31.9
Ian Heywood, Hans-RainerKl¨ockner, SteveRawlings: Equatorial Imaging with e-MERLIN + Chilbolton
1.0 1.0
0.5 0.5
C C
E E
S S
C 0.0 C 0.0
R R
A A
-0.5 -0.5
-1.0 -1.0
1.0 0.5 0.0 -0.5 -1.0 1.0 0.5 0.0 -0.5 -1.0
ARC SEC ARC SEC
Fig.4.Theresultofcleaningsimulated5minutesnapshotimageswithe-MERLINwithGaussiannoiseincluded.The
image with the Chilbolton antenna included is on the right. Contour levels are -30%,-20%, -10%,10%,20%, ... ,90%,
100% times the peak in the map (refer to Table 1).
1.0 1.0
0.5 0.5
C C
E E
S S
C 0.0 C 0.0
R R
A A
-0.5 -0.5
-1.0 -1.0
1.0 0.5 0.0 -0.5 -1.0 1.0 0.5 0.0 -0.5 -1.0
ARC SEC ARC SEC
Fig.5. Cleaned images of a model point source with the simulated observation being a full 12 hour synthesis using
e-MERLIN. The simulation with the Chilbolton antenna included is on the right. Contour levels are (–√2, 1, √2, 2,
2√2, 4, 4√2, 8, 8√2...) σ, where σ is the root mean square of the backgroundnoise (refer to Table 1).
×
array.Thisisparticularlyevidentinthecaseofequatorial hancement, such as high-resolutionradio surveys comple-
observations, where the dominant east-west layout of the menting optical and infrared deep-fields such as those to
existing array biases the uv coverage, resulting in strong beundertakenbyVISTA,andhigh-resolutionradioobser-
linear structure in the sidelobes of a synthesised beam vationsofgalacticandextra-galacticradiosourcestarget-
which is highly eccentric. ted with ALMA. Both e-MERLIN and ALMA naturally
complementeachotherdue totheirsimilarsub-arcsecond
Putting aside the increased sensitivity due to the
resolutions,despitetheiroperatingfrequenciesdifferingby
10% increase in collecting area with the additional an-
∼ a factor of 100.
tenna, the synthesised beam is much more compact and ∼
circular for equatorial observations, facilitating more effi-
Acknowledgements. The authors would like to thank Simon
cient detection of characteristic µJy radio sources, reduc-
Garrington of Jodrell Bank Observatory and Ken Craig of
ingthemappingspeedbyuptoafactoroftwo(seeTable2
Rutherford Appleton Laboratory.
for favourably oriented sources.
The addition of the Chilbolton antenna would signifi-
References
cantly enhance the power of e-MERLIN. There are many
experiments that would benefit significantly from this en- Muxlow, T.W.B., et al., 2005, MNRAS,358, 1159
Ian Heywood, Hans-RainerKl¨ockner, SteveRawlings: Equatorial Imaging with e-MERLIN + Chilbolton
Thomasson, P., 1986, RASQJ,27, 413