Workshop on Simulations for ALMA

[home]| [ALMA]


  • Alexandre Beelen, MPIfR/U Paris-Sud
  • Sandra Etoka, Manchester
  • Francesco Fontani, Geneva
  • Remy Indebetouw, UVa/NRAO
  • Brian Glendenning, NRAO
  • Frederic Gueth, IRAM
  • Antonio Hales, NRAO
  • Ian Heywood, University of Oxford
  • Mark Holdaway, Kalimba Magic
  • Eelco van Kampen, UIBK
  • Yasutaka Kurono, The University of Tokyo
  • Robert Laing, ESO
  • Robert Lucas, JAO
  • Koh-Ichiro Morita, National Astronomical Observatory Japan
  • Bojan Nikolic, Cavendish/University of Cambridge
  • Juan R. Pardo, CSIC
  • Dirk Petry, ESO
  • Jerome Pety, IRAM
  • Rob Reid, NRAO
  • Anita M. S. Richards, Manchester
  • John Richer, Cavendish/University of Cambridge
  • Nemesio Rodriguez, IRAM
  • Shigehisa Takakuwa, ASIAA
  • Francois Viallefond, l'Observatoire de Paris
  • Alwyn Wootten, NRAO
  • Thomas Wilson, ESO
  • Mel Wright, Berkeley
  • Mark Wyatt, IoA/University of Cambridge


This is the list of presentation already confirmed for the workshop. They are listed in order of arrival of abstract, not in order of presentations; the actual schedule will be published soon.

B. Nikolic (Cambridge): Simulating Atmospheric Phase Errors, Phase Correction and the Impact on ALMA Science

We present a framework for modelling atmospheric phase errors and their correction by both the fast-switching and water vapour radiometeric techniques. Notable features are simulating three dimensional turbulent volumes instead of flat phase screens, considering three representative configurations of ALMA and parametrisation in terms of phase fluctuations on a 300m baseline, allowing referencing to the site-testing interferometer data. We calculate relative point source sensitivity and effective resolution for a range of atmospheric conditions, with and without phase correction. We also consider the effect of phase fluctuation on short snapshot observations, where the variance in atmospheric phase fluctuations becomes important.

Shigehisa Takakuwa (ASIAA): Imaging Simulations for ACA Science

We are performing imaging simulations to assess the scientific role of ACA. Our recent observations of several low-mass protostellar envelopes in the submillimeter CS (J=7-6) and HCN (J=4-3) lines with the SMA and ASTE have revealed that these submillimeter emissions are more extended than ~ 2000 AU and show different velocity structures from those traced by millimeter lines. These results suggest the importance of taking short-spacing informations ACA can offer. Our comprehensive imaging simulations of these protostellar envelopes, as well as prestellar cores and debris disks, unprecedentedly demonstrate the scientific importance of ACA.

R. Indebetouw (NRAO): The CASA ALMA Simulator

I will demonstrate the current capabilities in CASA for simulating ALMA (and other array) data. I will also discuss our near future development plans. CASA will have the capability to generate realistic ALMA measurement sets of channelized mosaic observations, degraded by thermal noise, atmospheric phase noise, pointing and other errors. Various metrics such a s fidelity comparing a model and synthetic image can be automatically calculated. One will be able to create datasets containing science target, calibrator, and WVR information and seamlessly use the offline calibration tools in CASA to assess the entire data analysis process.

Mark Wyatt (IoA): Debris disks at high resolution

Some 15% of nearby stars are surrounded by dust that derives from planetesimal belts that are analogous to the asteroid and Kuiper belts of the Solar System. So far 20 of the nearest and brightest of these debris disks have been imaged from optical to millimetre wavelengths. These images show that the disks are highly structured, containing asymmetries and clumps that can be attributed to perturbations from an unseen planetary system. High resolution imaging with ALMA presents two exciting opportunities: (i) to map the structure of known disks with a resolution that will be able to distinguish between different models for the structure formation, and (ii) to resolve the structures of a significantly larger number of disks. This talk will present simulations of ALMA observations of debris disks and discuss the ways in which such observations will constrain our knowledge of the planetary systems of nearby stars.

Anita Richards (Jodrell Bank): Virtual Observatory publishing of interferometry simulations

Simulated data for ALMA, SKA, e-MERLIN or any instrument can be shared using the Virtual Observatory and compared with existing observations at any wavelength. In its simplest form, this involves publishing a static source list using a database linked to web services provided by a VO. Images, spectra etc. can be provided by a similar method. In all cases, the data are described using International Virtual Observatory standards so that they can be retrieved and manipulated using any VO-compatible tools all over the world (access restrictions are also possible). It is possible, however, to go much further and generate simulations dynamically, on request, via a simple user interface (typically a few dialogue boxes or menus, or a short script). The AstroGrid UWS transmits the request to a pipeline or any command-line program which performs the necessary operations and returns the results. If these are bulky, the response can be in the form of a descriptive table containing the URLs of the actual images (etc.) which are only downloaded on request. Similar services for 'real' data are already in operation based on the RadioNet parseltongue AIPS wrapper. We will demonstrate some possibilities and explain how to get further support from your local VO or the EuroVO DCA.

Kurono, Morita, Kamazaki: Imaging Simulations and Analytical Consideration for the Technique of Combining Single-dish and Interferometer Data

We conducted imaging simulations of the data combining of a single-dish telescope and an interferometer, on the assumption of the combination of NRO 45m Telescope and Nobeyama Millimeter Array at Nobeyama Radio Observatory (NRO). We demonstrate the effects of data combining, and discuss principal issues that have not been clarified up to now: (1) relative weight between the interferometer data and the single-dish data when the data combining is performed, and (2) sensitivities in each observation for the data combining, and an analytical treatment of them. Our approach and consideration can be generally applied for ALMA with ACA system in the near future.

N. Rodriguez: Simulations and processing of ALMA On-the-fly observations

One of the most promising observing modes foreseen for ALMA observations of extended sources is the On-The-Fly (OTF) observing mode. The OTF mode is not only less time-consuming than classical stop-and-go mosaics, it will also allow to observe wider fields and it also opens the way to new methods of data editing and processing. At IRAM, we are developing OTF specific imaging and deconvolution algorithms. As a first step we have developed an OTF simulator based on the IRAM/GILDAS ALMA simulator. In this contribution I will present the project and the current state of the work.

J.Pety (IRAM), F.Gueth (IRAM), S.Guilloteau (Obs.Bordeaux): The ALMA imaging simulator in GILDAS

GILDAS, the software suite developed at IRAM, includes an ALMA imaging simulator. It has been developed a few years ago to make extensive tests of the impact of the ACA on the wide-field capabilities of ALMA (see e.g. ALMA memo 398). Since then, it has been used for further instrumental simulations (e.g. ALMA memo 488) and for the scientific preparation of ALMA (e.g. Wolf et al. 2002, 2005, Semenov et al. 2008)

From an input FITS file, this tool simulates observations, including mosaicing and short-spacings, then performs the imaging and deconvolution of the data. Various errors can be included, such as atmospheric phase noise (using a phase screen), thermal noise, pointing errors, calibration errors. Different error models can be selected, e.g. coherent pointing errors on all antennas [typical for sunset/sunrise], fully random errors, drifts, etc. A comparison with the input image is performed by computing fidelity distribution and various image quality indicators.

In this talk, this simulator will be presented, as well as the further improvements that are planned or desirable.

F. Viallefond (LERMA) and Juan Pardo (DAMIR): Status of and perspective for tools used to simulate ALMA-like data

We developped a simulator, SIMULATM, about 4 years ago. This tool has been used primarily to test software components that we developped in the computing IPT for the TelCal and OFFLINE subsystems. At that time we developped an interface to the ATM library which was intended for client applications using the Fortran and C language.

SIMULATM was also used to create samples of datasets to test data models, in particular the Alma Science Data Model (ASDM). For that purpose we had developped a C to C++ interface.

Since that time both ATM and the ASDM evolved very significantly. ATM has been totally rewritten in C++. The ASDM evolved with, in particular, the definition and implementation of the ASDM Binary Data Format (BDF) and the Data Access Method (DAM) for the data produced by various processors. Furthermore a set of tables were added to host the TelCal results. A new strategy recommended by the project (UML/OAW) had also to be used to generate the code for the ASDM.

The main body of SIMULATM being written in C it became difficult to keep track of these evolutions. In that context totally new software components were developed to generate ASDM samples of datasets with their associated Binary Large OBjects (BLOBs). These are used to finalize the ASDM, the BDF and the DAM for the ASDM-to-MeasurementSet filler. They are also critical for our work about the definition and implementation of the Export Data Format (EDF). In these components most of the attention has been paid on the data structure and data flow, the actual values for the data produced by the different processor being not important.

Hopefuly, with the inclusion of the physical quantities and measures in the ASDM it will be easier to develop application using the ASDM and CASA. A model has been developped and prototyped for that purpose. At that point the ASDM should be sufficiently stable to resume with the development of a tool like SIMULATM to simulate ALMA observations.

SIMULATM had the following functionalities: - simulation of an atmospheric phase screen moving above the array - simulation of the WVR data with one radiometer per antenna, each with its own characteristics (stability, ...) - phase errors in the signal path - receiver thermal noise - phase correction using the simulated WVR data - phase referencing with fast switching

SIMULATM was functioning with a several servers for - configuring the antenna positions (up to 64) - configuring the correlator setup with 4 basebands and Tunable-Filter-Banks (TFBs). - the production and evolution over time of the phase screen - receiver 1/f noise generators - capturing the simulated data in a MS (version pre-ASDM) and ALMA-TI-FITS

SIMULATM is an event-driven application which is controlled either interactively or via a script. Its GUI includes graphics for monitoring the phase the phase correction, the simulated phase structure function, the image of a point source generated from the simulated visibilities radiometricaly corrected or uncorrected.

Perspective: the plan is to re-write totally SIMULATM - with a single language, C++ to make it simpler to maintain - using the C++ interface of ATM - with the necessary upgrades for missing instrumental effects (ref SSR doc.) - to produce dataset always supporting the most recent version of the SDM - possibly to allow to run it with standard ALMA observing scripts.

Eelco van Kampen (UIBK): Mock maps of the high-z universe: from SCUBA-2 and Herschel to ALMA

Currently simulators are concentrating on the upcoming large surveys with SCUBA-2 and Herschel, which have comparable resolving power. Methods will need to be improved for simulating ALMA maps, though. Its resolution is so much better that the modelling of the high-z galaxy population will need to be much refined, especially with respect to groups and pairs of galaxies, which SCUBA-2 and Herschel cannot resolve at high redshifts. ALMA will be able to shed (dusty) light on this, but much better simulations will be essential.

Rob Reid (NRAO): Evaluating the extended ALMA configurations using almasimmos

ALMA's 20 most compact configurations are spirals occupying a roughly circular plain, and they will provide excellent uv coverage. The eight more extended configurations, however, are constrained by the terrain to resemble a Y, although with less regularity than the VLA configurations. This restriction necessitates some compromises on the uv coverage. Lewis Knee and I have recently completed evaluations of whether the resulting uv coverage is sufficient, after deconvolution for ALMA's needs, and whether some modifications of the configurations would be better. I will also discuss measures of configuration and imaging quality, and what features of a configuration seem to really matter.

Al Wootten (NRAO): ALMA Becomes Real

(no abstract at this time)

Mark Holdaway (Kalimba Magic)

Earlier simulations comparing simulations with 50 and 64 antenna ALMAs did not indicate that 64 antennas provided a large improvement over 50 antennas. At this time, we look into a more realistic problem: how do 50 antennas compare with 64 antennas when we need to solve for the source structure along with other (self) calibration variables such as atmospheric phase and amplitude, and pointing errors?

Ian Heywood (Oxford): ALMA and the SKADS Simulated Skies

I will be presenting an overview of the aspects of the Oxford simulations effort which will be of interest to the ALMA community.

We have developed a simulated radio sky covering 20 x 20 square degrees out to a redshift of 20 with a 10 nJy flux limit and which includes both continuum (Wilman et al, 2008) and spectral lines (Obreschkow et al, in prep). I will present some simulated ALMA observations based on a high-redshift CO source selected from this database.

We have also been employing a raytracing technique to simulate the effects of gravitational lensing due to rich foreground clusters, and I will demonstrate the effect of having an Abell 2218-like cluster in front of a high redshift CO source when being observed with ALMA.

Finally, I will discuss the advantages offered by using MeqTrees as a simulations package. The MeqTrees software has been recently paralellised at the Oxford e-Research Centre, opening up the possibility of simulating densely populated skies with large-N arrays, and moving from the idealised case to a plausible one.

The open-ended nature of MeqTrees also allows features such as tropospheric models, and both image- and Fourier-domain corruption effects to be readily incorporated into simulations.

Melvyn Wright (Berkeley) & Stuartt Corder (Caltech): Deconvolving Primary Beam Patterns from Mosaic and Polarization Images

We present a method for deconvolving the primary beam response from interferometric images of astronomical sources. The measured primary beam may be time variable, non axi-symmetric and different on each antenna in the interferometer array. The method is a simple extrapolation of existing software which subtracts a model of the sky brightness distribution from uv data. After subtracting the best estimate of the sky brightness distribution weighted by the measured primary beam pattern, the residual uv data can be re-imaged to provide an improved model of the sky brightness distribution and the process iterated if needed until the residual uv data are consistent with thermal noise and other residual instrumental errors. The data are imaged using canonical, time invariant primary beam patterns, and deconvolved using the measured primary beam voltage patterns for each antenna.

We simulate observations with the CARMA telescope and calculate the errors which result from the measured primary beam voltage patterns at 100 GHz. We show that the effects of the measured ~ 1 - 5% deviations from the canonical beam patterns can be devastating, reducing the image fidelity from ~ 8000 to ~ 50 for a source which fills the primary beam FWHM. The image fidelity can be greatly improved by using the measured voltage patterns in the deconvolution.

The primary beam pattern is the product of the voltage patterns for each antenna pair, and is complex valued if the voltage patterns are not identical. This results in a complex valued image of a real, total intensity, sky brightness distribution; i.e. the image shows a polarized flux distribution which varies across the primary beam, and any real polarization distribution is confused by flux scattered from the total intensity by primary beam errors. Polarization images can be corrected by subtracting a model of the source weighted by the complex valued primary beam patterns from the uv data.

Robert Lucas (JAO)

ALMA Shared Simulator