Session I

13:30 - 15:50

13:30

Huib Jan van Langevelde

ALBUS objectives

14:05

Mark Kettenis

parselTongue: AIPS talking Python

14:40

Joe McMullin

Offline Analysis Software for ALMA and the NRAO EVLA

15:15

Steven Myers

EVLA Data Post-processing Plans and Algorithm Development

Session II

16:20 - 18:30

16:20

Garrelt Mellema

Parallelization Lessons from Miriad

16:55

Harro Verkouter

Data flow Beyond PCint

17:30

Discussion

Status of coding platforms

18:30

Finish

Session III

09:10 - 10:35

09:10

Tom Oosterloo

Correcting for image plane effects in WSRT data

09:45

Jan Noordam

Non-isoplanatic patch effects

10:20

Ger van Diepen

Implementation of LOFAR calibration

Session IV

11:05 - 12:15

11:05

Huib Jan van Langevelde

Wide-field imaging with the EVN

11:40

Vincent McIntyre

Multi-Frequency Synthesis: developments at ATNF

Session V

13:30 - 15:15

13:30

Anita Richards

Publishing radio data to Virtual Observatories

14:05

Vincent McIntyre

Virtual observatory work at ATNF with the aips++ toolkit

14:40

Friso Olnon

EVN archive

Session VI

15:45 - 17:50

15:45

Discussion

Status of coding platforms (continued)

Wishes for future software improvements

17:50

Finish

Session VII

09:15 - 10:25

09:15

Cormac Reynolds

Calibration transfer for the EVN

09:50

James Anderson

Ionospheric calibration for the EVN

Session VIII

10:55 - 12:30

10:55

Roy/Rottmann

Tropospheric calibration for VLBI

11:30

Arpad Szomoru

eVLBI

12:05

All

Summary/Conclusions/Action Items/Next Meeting Plans

Private Session I

ALBUS management meeting

14:00

Huib Jan van Langevelde et al.

16:00

Finish

TITLES/ABSTRACTS RECEIVED

James Anderson

Ionospheric calibration for the EVN

As part of the ALBUS project, we are evaluating and implementing several algorithms to apply different ionospheric correction methods to VLBI data. We are focusing on methods to predict the total electron content along the observed slant paths for each VLBI antenna. These methods will include estimates based on dual-frequency GPS measurements of the ionospheric delay from GPS receiver stations co-located with VLBI antennas. Empirical and theoretical ionospheric models (IRI, PIM), and two and three-dimensional global ionospheric electron content models (e.g. the IONEX global vertical TEC files and Fusion Numerics' numerical ionospheric forecasting system) will also be included for analysis. We plan comparison tests for the predicted slant TEC values from these various data sources for actual EVN observations. The amount of coherence gain provided by the different correction methods will be evaluated for different observing times, elevations, and phase referencing configurations.

Ger van Diepen

Implementation of LOFAR calibration

The LOFAR instrument can produce large amounts of data (a few TBytes per hour) which means that highly distributed data storage and data processing is needed in order to calibrate and image an observation in a reasonable amount of time. The first part of the calibration is done in the online processing pipeline. The final calibration and imaging is done in an offline process. The control of the system is done by means of a Blackboard architecture.

Mark Kettenis

parselTongue: AIPS talking Python

ParselTongue is an environment that is intended to make it possible to integrate existing Data Reduction Packages (in particular classic AIPS) with Python. Currently its primary focus is on making these packages scriptable from Python. However it is designed with interactive use in mind. It allows for executing tasks remotely (currently implemented using XML-RPC) in addition to the traditional method of local execution. This allows for course-grained parallelisation on the task level. ParselTongue also provides access to (meta)data; an essential feature for implementing non-trivial pipelines. For this purpose it uses functionality from Bill Cotton's Obit. ParselTongue and Obit will (hopefully) provide the infrastructure necessary for implementing the advanced calibration algorithms that are developed as part of the ALBUS project. I will start my talk by giving a short introduction about Python and its use in astronomy. I will then continue to give an overview of the capabilities of the current ParselTongue system and I will explain some of the interesting implementation details.

Huib Jan van Langevelde

ALBUS objectives

The ALBUS (Advanced Long Baseline User Software) is an EU supported software project that involves partners at four institutes. It covers several aspects of the dataflow for radio-interferometry and especially VLBI. An effort will be made to improve the calibration of the EVN, not only by making the calibration more transparently available, but also by providing estimates for tropospheric and ionospheric calibration. Another area is image processing with the enormous data volumes that new instrumentation can produce. A special issue has been to decide on which platform to develop new software.

Huib Jan van Langevelde &
James Anderson

Wide field imaging with the EVN

New hardware for the EVN and eMERLIN will allow large fields of view to be imaged with a single data set. As these large fields of view require fine sampling of the output visibilities in frequency and time, the resulting datasets are large by current standards. We discuss the existing techniques for wide field imaging with respect to the special case of long baseline interferometry. Some preliminary ideas on paralellization of these algorithms will be presented.

Vincent McIntyre

Virtual observatory work at ATNF with the aips++ toolkit and Multi-Frequency Synthesis: developments at ATNF

ATNF has been successfully reusing elements of the AIPS++ toolkit on projects in both the virtual observatory arena and for our telescope support requirements. I'll describe three software projects that have relied on this codebase - a package for processing single-dish spectra, a system for server-side visualisation of data holdings and an automated processing pipeline for archival data - and highlight how the strengths and weaknesses of the toolkit affected these development projects.

Joe McMullin

Offline Analysis Software for ALMA and the NRAO EVLA

The AIPS++ Project developed a code base of libraries, toolkits, and applications for the analysis of radio astronomical data. This code base has been reworked into a layered set of libraries (CASA - Common Astronomy Software Applications) which are being used for the development of the offline analysis software for ALMA and the EVLA. We discuss the ongoing testing of applications (validity, robustness, benchmarking), algorithm development (e.g., W-projection, antenna-based pointing errors, wide band imaging) and the use of ACS (ALMA Common Software) as the execution framework for a task based offline analysis system.

Garrelt Mellema

Parallelization Lessons from Miriad

Steven Myers

EVLA Data Post-processing Plans and Algorithm Development

NRAO is in the process of formulating the plan for the post-processing of EVLA Phase I data. This involves software development and implementation in AIPS and aips++ (co-developed with ALMA), as well as extensive algorithm development. The hard problems for EVLA I are wide-field, wide-band (e.g. 2:1 bandwidth), high-fidelity, and high-dynamic range imaging, particularly at low-frequencies where ionospheric distortion is a signficant effect, and it is clear that there needs to be significant research into new imaging and calibration techniques in order to meet the ambitious goals for the project. The software and algorithms are also being developed in concert with LWA project, where these issues are even more severe. I will discuss our current thinking on the software platforms (both for the code base and the possibility of a unified framework under which the code could be run) and development plans, describe our algorithm research effort, and the possibilities of co-development and collaboration with similar efforts underway for other projects in the community.

Jan Noordam

Non-isoplanatic patch effects

The MeqTree system (formerly known as the Prototype Selfcal System, or PSS, for LOFAR) is a module that can implement an arbitrary Measurement Equation (M.E.), and solve for subsets of its parameters. As a module, it makes use of modules from other packages, especially AIPS++. An overview will be given of the module, and its use for some of the new algorithms like 'peeling'.

Friso Olnon &
Bauke Kramer

EVN Data Archive at JIVE

The EVN Data Archive is now operational for two years. It contains the user data for all 200+ experiments correlated at JIVE. That amounts to 0.8 TeraBytes, mostly in the form of FITS files. The archive is filling up quickly, and expansion from its present 2 TB to 10 TB is in progress. The access procedures have extensively been used by JIVE scientists and visiting P.I.'s. No serious problems were encountered, and new features and tools are continuously added on request by the users. The archive is not often accessed by 'outsiters', because we cannot offer them very useful information and/or because data transfer rates are fairly low. We are working on improvements in both areas.

Tom Oosterloo

Correcting for image plane effects in WSRT data

The standard way of calibrating radio interferometric data is to assume that errors are antenna based, and the one correction factor per antenna (the gain) for the entire field is sufficient. However, pointing errors, small difference between dishes as well as ionospheric effects, make this assumption invalid, even for single field imaging at 21 cm with the WSRT. For high-dynamic range imaging, a more elobarate approach is needed. The technique I have explored is to derive different gains for different parts of a single image in a process similar as is being developed for Lofar. The multi-gain approach can lead to dramatic improvement in the dynamic range of the images. I will show application of the technique to several WSRT 21-cm data sets.

Cormac Reynolds

Calibration transfer for the EVN

Much of the calibration required for VLBI arrays is achieved with the use of auxiliary data which are not directly incorporated into the visibility products themselves. For data from the EVN correlator at JIVE, these auxiliary data are currently provided in the form of formatted text files. This system has a number of drawbacks, especially in terms of data archiving and retrieval and is at least partly responsible for the perception of VLBI as being difficult" for non-experts. A simple solution to this problem is to attach the auxiliary calibration data to the distributed FITS file such that it appears as an attached calibration table as soon as the data is loaded in the relevant data analysis package (this is known as calibration transfer). I will present a progress report on efforts to implement this calibration transfer scheme at the EVN correlator at JIVE.

Anita Richards

Publishing radio data to Virtual Observatories

I will report on various routes to data of different types (images, visibilities, spectra, catalogues), what Virtual Observatories offer at present, what radio archives need to do and suggestions for future developments by both parties. I will also touch on the contribution of VOs to support of designing observations (such as calibrator searches and the process of setting international standards). Finally I will demonstrate recent science applications of VO tools, including using radio data.

Helge Rottmann &
Alan Roy

Tropospheric calibration for VLBI

A new 22 GHz water-vapour radiometer was installed during 2003 at the Effelsberg 100 m telescope, for tropospheric phase and delay and opacity correction during high-frequency VLBI observations. This adds to the small but growing number of EVN and millimetre telescopes that are equipped with radiometers. We will describe the radiometer, the data path being opened up to the correlator and expected user scenarios.

Arpad Szomoru

eVLBI

Harro Verkouter

Data flow Beyond PCint