WP 1.4 Coordination of LSST:UK Contributions to Commissioning

Issued to the UK community an EoI Call for Rubin Observatory Commissioning. An accompanying briefing paper aims to assist colleagues in preparing for and contributing to commissioning. This is in the context of the current pause in Rubin construction, and the ongoing negotiations of the UK’s in-kind contributions to the Rubin Observatory and LSST Science Collaborations.  This call closed on October 30^th 2020.  

WP 2.1 DAC Management

Data Access Centre requirements have been integrated into the LSST:UK Science Requirements Document (Version 3.0, April 2020).

WP2.2 Data Ingestion and Publication

A set of data transfer and ingest experiments were undertaken, between IN2P3 (the Rubin data-processing facility in France) and the Royal Observatory Edinburgh. These are documented in D2.2.1. Following on from this, work is underway to ingest ancillary data for the next iteration of the DAC (UKDAC1), including ingestion of PanSTARSS DR2 and ZTF DR2.

WP2.4 Provision of the DAC Platform

The DAC team has begun collaborating with Rubin Observatory staff in the United States and France around the development of the Rubin Science Platform (RSP). WP2.4 staff are customising the baseline RSP for UK-specific ancillary datasets and IRIS IAM authentication, towards the roll-out of an updated RSP in the UK DAC in 2021Q1.

WP2.5 Science Support

Deliverable D2.5.1 “Training resources for LSST:UK DAC users” was completed. This describes an initial release of documentation for users of current and future services accessed via the UK’s LSST Data Access Centre (DAC). This documentation release is necessarily limited in scope given that the UK DAC is still being developed. It comprises existing documentation for the Lasair alert broker and a very preliminary set of documentation for the LSST Science Platform (LSP). The LSP is the set of data services to be provided by the Rubin Observatory to support analysis of LSST data products.

WP2.3/3.2 Lasair

The team has: completed and released Lasair v2.0; tested the Cassandra database architecture on IRIS and reviewed it as a future implementation for LSST scale alerts (Report in preparation at the time of writing); completed a science requirements and functionality review led by the LSST:UK PoCs; tested the RAPID light-curve classifier and published a technical summary paper.

WP3.5 LSST and near-infrared data fusion

The WP team has copied over onto the IRIS HPC infrastructure all publicly available VISTA imaging survey datasets and work is now ongoing to develop the pipeline to process the VISTA pixels through the Rubin stack. Current efforts in the WP are focused on understanding photometric calibration issues as well as better understanding the noise properties of the resulting images produced by the pipeline. The current plan is to have a test region in the SXDS field processed and available for scientific validation by December 2020.

WP3.7 Low-surface-brightness science using LSST

Team member Aaron Watkins has been appointed as Deputy Chair of the LSST Low Surface Brightness Working Group. He is working alongside Sarah Brough from the University of New South Wales Australia to help coordinate international community effort in infrastructure development related to intra cluster light and dwarf galaxy science. This includes, although not restricted to, activity related to international LoI proposals. Aaron also gave an update of his WP 3.7 work at the LSB Session of the Rubin Project and Community Workshop in August. He presented results quantifying the over subtraction in different versions of the current data pipeline to an audience of well over 100 participants. Aaron has also given virtual seminars at LJMU and Hertfordshire and to the LSST Galaxies forum and the LSB Challenge 1 meeting, of which he is co-chair.

WP3.9 LSST Point Spread Function, sensor characterisation and modelling

This WP team have established the following.

• Confirmation of optimal gate width settings in e2v CCDs – this WP were working on finding what setting of the CCD gate width parameter was optimal from the perspective of correlation and full well effects. The investigation has shown that one particular setting outperforms all others in most reasonable scenarios, and by coincidence this was the setting already used by LSST camera.

• Confirmation that silicon di-vacancies cause trap-related parallel CTI in e2v CCDs – investigations over the last year have produced strong evidence of the particular type of trap species that dominates in the operating conditions used in the LSST camera. From this, timing recommendations can be made to substantially reduce parallel CTI, an important performance degrading effect. Other CTI mechanisms than trapping exist in these devices and it is not yet fully clear what proportion of CTI is caused by trapping, but in a traditional CCD it is the dominant effect.

• Calibration Improvements in Oxford LSST testbench – due to renewed efforts to carefully track down systematics and calibration errors in our lab equipment and testbench, and thanks to the help of very strong undergraduate project students, this WP have been able to reliably reduce the illumination fluctuations in experiments performed on the OPMD lab testbench. This will aid in future WP efforts that require much more stable operation than was previously possible.

• Evidence that previous assumptions used in the theory of charge trapping in CCDs are incorrect – the WP have noted for quite a while that previous analyses of charge trapping in CCDs contained a somewhat unjustified assumption in the derivation of the equation governing trapping rate that essentially only applies to traps which happen to lie in high charge density areas of the device. Since traps are uniformly distributed in space this assumption does not hold in reality. The WP have made significant progress in developing an analysis method that relaxes this assumption. Results from the WP new analysis method are so far consistent with previous methods, yet have two advantages: a) they pave the way in future to use the same data to extract more detail about charge density than was previously possible, and b) they reduce the unexplained variance in individual trap energy levels that have been measured with the old methods

WP3.10 UK Contributions to DESC Operations

Deliverable D3.10.5 “Processing DC2 data using the LSST DM Stack on UK Facilities, was completed. This deliverable describes how to process the images generated by DESC's Data Challenge 2 using the LSST software pipeline. Instructions for installing the software, setting up a data repository and running each stage of the pipeline are provided. Possibilities for using container and workflow technologies to improve the process are also discussed.

WP3.11 Cross matching and astrometry at LSST depths

The WP team completed their first deliverable, D3.11.1, an investigation into the model for contamination of sources due to crowding at LSST depths. WP 3.11 are confident that they can model the astrometric perturbations and photometric contaminations of sources due to faint, blended objects within their PSFs and have algorithms in place that improve the accuracy of the simulated models across a wide range of signal-to-noise ratios, from bright, photon-dominated objects to (important for LSST) faint, sky background-dominated objects. In addition, this WP have started building the software framework for fully symmetric, many-to-many Bayesian cross-matches. This code will serve as the preliminary test bed for investigations in integrations with the DAC workflow, user interaction with the end products, profiling to LSST data sizes, and an analytic groundwork from which to probe reproducibility, recovery rates, etc.  The WP are also are working with the LSST:UK DAC team to establish a “data challenge” with the preliminary codebase. This involves a full end-to-end test — using Gaia and WISE, with WISE serving as a proxy for LSST in sources-per-PSF space — to verify and test a full-scale DAC-DEV product creation. Finally, this WP have also begun discussions with various US-based teams — both the DM team within LSSTC itself, and science collaborations — on coordinating efforts on understanding the LSST datasets. Links have also been made with the TVS SC, with potential collaboration efforts on the cross-matching of LSST alerts in real-time, the SMWLV SC, focussing on the importance of our cross-match algorithms in crowded Milky Way fields, and the Crowded Field Task Force in the LSSTC DM, again linking the WP 3.11 cross-match expertise but also working with them to ensure that the WP 3.11 implementation uses as robust and precise a description of the LSST data processing as possible.

D2.3.1 Lessons learned from ZTF

Roy Williams, Ken Smith, Stephen Smartt, Andy Lawrence, Gareth Francis

The LSST:UK project has built Lasair, (https://lasair.roe.ac.uk), a community broker for the LSST alerts that will encourage a variety of users in a variety of science investigations. The objective is twofold: to give UK and other scientists access to the LSST stream in near-real-time, and also to add value to those alerts. In 2018, a web-database prototype broker for the Zwicky Transient Facility (ZTF) was built, that has a similar structure to what is expected for LSST, using Kafka to deliver alerts to their brokers. Both LSST and ZTF send an alert whenever a source is significantly (5σ) brighter than it was in an earlier reference sky. This paper reports on experiences of running this prototype.