2026-11-02 –, Conference Foyer 2
The CASA/ALMA Pipeline is the standard data processing software for the Atacama Large (sub-)Millimeter Array (ALMA) radio interferometer. It has been developed by an international team over the course of more than 20 years and it is now used to process >95% of all ALMA datasets.
The planned ALMA Wideband Sensitivity Upgrade (WSU) will be enhancing the observing capabilities with larger bandwidths and better spectral resolution, leading to orders of magnitude higher data volumes which will need a new highly parallelized data processing system.
In this poster we describe ongoing work to prepare the pipeline for this future workflow based system. We discuss refactoring the interferometry imaging, the most time consuming step in the pipeline, to use faster characterization of features in datasets via region-of-interest detection, reducing pre-imaging workloads and mitigation strategies while limiting any scientific loss.
We are also working on analyzing the so called "Pipeline Context" which acts as a central hub to record stage results and to transport intermediate products for use in later stages. Its current mixture of use cases needs to be modularized and distributed to be scalable.
To enhance ALMA's "Fast Lane" delivery of pipeline results, this cycle we have worked on improving a number of our Quality Assessment (QA) Scores to allow for more robust identification of pipeline runs that can be immediately delivered to the PIs without detailed human inspection.
Finally, a large documentation restructuring effort using "readthedocs" has been undertaken to streamline and improve existing documents and facilitate adding more content which will help transfer the existing pipeline to the new data processing system.
Staff scientist at MPIfR, Bonn, Germany. ALMA pipeline developer since 2003. Also in charge of the APEX telescope control and calibration software. Working on atmospheric models. Interested in ISM turbulence.