The technical capabilities of e-MERLIN
MERLIN is an array of seven radio telescopes across the UK, connected to a central correlator at Jodrell Bank Observatory (JBO) and operated as a dedicated radio interferometer to produce high-resolution images. With a maximum baseline length of 220 km, MERLIN provides a unique capability for radio imaging at 0.01-0.15-arcsec resolution at frequencies of 1.5, 5 and 22 GHz (L, C and K bands). The e-MERLIN project is a major upgrade to the instrument involving the installation of new receivers, analogue and digital electronics, optical-fibre links to each telescope and a new correlator at JBO. This will increase the useable bandwidth by more than two orders of magnitude, and hence the continuum sensitivity by more than 10x. In addition, the increase in bandwidth will dramatically improve aperture coverage for continuum observations resulting in enhanced image fidelity together with simultaneous spectral-index imaging. The e-MERLIN upgrade is now well underway: new receivers are in service, the optical fibre network has been installed and the digital transmission equipment has been tested. First fringes between nearby telescopes using a prototype correlator were achieved in April 2009, fringes from all remote telescopes were obtained in autumn 2010, and the full network with the complete correlator will be available in 2012.
Technical Capabilities - FINAL ARRAY
Table 1: Basic observing capabilities of e-MERLIN
|Resolution (mas)||150||40||12||Uniform weighting at central frequency|
|Field of View (arcmin)||30||7||2.0||FWHM of 25-m dishes; reduced when Lovell Telescope included at 1.5 or 5 GHz (1)|
|Freq. Range (GHz)||1.3-1.7||4-8||22-24|
|Bandwidth (GHz)||0.4||2||2||Max. Bandwidth per polarization. Can use 4-GHz, single polzn, at 5 or 22GHz|
|Sensitivity (μJy/bm) in full imaging run||5-6||1.8-2.3||~15||Final performance will depend on useable bandwidth, final reciever optimization, Lovell Telescope performance. These figures are for e-MERLIN with the Lovell Telescope(1).|
|Surface brightness sensitivity (K)||~190||~70||~530||As above|
|Astrometric performace (mas)||~2||~1||~2||WRT the ICRF (typical 3-deg target-calibrator separation using VLBA Calibrator Survey)|
|~0.5||~0.2||~1||Day-to-day repeatability using surveyed or in-beam sources, and assuming full imaging run.|
|Amplitude calibration||2%||1%||10%||Targets for day-to-day repeatability|
Notes: (1) The Lovell telescope may be included in e-MERLIN at 1.5 and 5 GHz (L, C). Its inclusion increases the sensitivity by a factor of between 2 and 3 and reduces the field of view to approximately 20/(freq/1.4GHz) arcmin, depending on the data-weighting scheme adopted.
The new e-MERLIN correlator, being developed by DRAO, will provide unsurpassed capabilities with very large numbers of configurations - Tables 1 and 2 which illustrate some of these. Further, narrower bands are also available. 3-bit sampling is standard at 5 and 22 GHz (C, K). 8-bit sampling is the preferred choice at 1.5 GHz (L) where the total available bandwidth is less and where the RFI environment is likely to be significantly worse. The centre frequencies, bandwidth and frequency resolution of each sub-band can be set individually, allowing a wide variety of mixed spectral-line and wide-band modes. Sensitivities per channel in spectral line modes can be estimated from the continuum sensitivities given above and by scaling with (δν/Δν)^1/2 where δν is the channel width and Δν is the total bandwidth: 0.4 GHz at 1.5 GHz (L), or 2 GHz at 5 or 22 GHz (C, K). Sub-bands can be placed anywhere within integer sub-slots corresponding to the sub-band width. Recirculation will produce very narrow channels but potentially very large datasets. For continuum modes, the quoted field of view is the diameter assuming 10% radial smearing at the field edge, but other effects may dominate beyond the half-power point of the telescope primary beam.
Table 2: Correlator capabilities of e-MERLIN
|Bandwidth (MHz)||Bits Sampling / correlation||Sub-bands||Sub-band bandwidth (MHz)||No. of Polzn||No. of channels
(per polzn, per sub-band)
|Channel width (kHz)||Field of view (arcmin)|
|5- and 22-GHz (C, K) continuum modes:|
|1.5-GHz (L) continuum modes:|
|Spectral-line modes (maximum recirculation, 4-bit correlation):|
|Spectral-line modes (maximum recirculation, 7-bit correlation):|
|Spectral-line modes (maximum no-recirculation, 4-bit correlation):|
Notes: (2) The field of view for spectral line modes is ultimately limited by the telescope primary beam but in some cases it may be limited by the output data rate
Operational opportunities and constraints
e-MERLIN will be dynamically scheduled in order to optimize the efficiency of full track observations, making the best use of observing conditions, and fitting in low priority observations when one or more telescopes are unavailable. Specific scheduling for monitoring purposes, or to catch particular phases of target source behavior can be accommodated, but will have an increased impact on the overall observing efficiency. Surveys requiring short observations of large numbers of targets in individual pointings are feasible but may require careful optimization. Deep field observations (especially at higher declinations) can make good use of periods when one telescope is unavailable, in order to build up total integration time.
It will be possible to attain even higher resolution with joint e-MERLIN/EVN observations, or include large-scale structure in images by combining e-MERLIN and EVLA data. This will be much simpler than in the past thanks to the new, flexible correlator.
Data handling and processing issues
We anticipate that e-MERLIN data will be processed using a combination of AIPS, casa and other interferometry data-reduction packages, using scripts which may also allow inter-operability between different packages. Raw data volumes of big projects are likely to be large (up to TBs depending on programme size and configurations) and the data will be distributed on hard disks. Data will be made available as UVFITS files with associated calibration tables. Options for making data available in other formats (casa Measurement Sets; ALMA/EVLA SDM and binary data) are under consideration.
We are developing scripts and pipelines to improve the accessibility of data products and images, once e-MERLIN is fully operational.