This month saw about a hundred of us gather in Manchester for an LSST:UK All Hands Meeting, held in conjunction with Euclid UK: two days of LSST-specific sessions, followed by a day on LSST-Euclid synergies, prior to a two-day Euclid UK Consortium meeting. Presentation material from the first two days is available here, while summaries of the LSST-Euclid discussion sessions are currently being prepared. We will shortly be circulating a brief questionnaire to all attendees to gather feedback on this meeting and to aid planning for future events. Thanks to the AHM SOC – Graham Smith (chair), Rebecca Bowler, Chris Frohmaier, Andrina Nicola, Cyrielle Opitom, Clara Pennock, Heloise Stevance, Matthew Temple and Tom Wilson – for putting together such an enjoyable programme and to all participants for engaging with it so enthusiastically; our most particular thanks go to our local hosts, Rebecca and Andrina, for logistical planning over several months that made it all possible.

We were delighted that the Rubin Director, Bob Blum, was able to join us remotely and present an update on the continuing Early Operations optimisation work that provided more details on many of the topics that have been covered by Keith Bechtol in his weekly updates on the Community forum. Bob noted that a major update will be presented to the community at the end of April, by which time Rubin staff will have digested the wealth of information they have obtained from the recent weeks of on-sky observations assessing the variety of factors that are affecting the reliability of delivered image quality.

We have been allocated three parallel sessions, plus a lunchtime session for the Junior Associates Network (see below) at NAM2026. If you would like to join the Scientific Organising Committee for the parallel sessions, please email Graham Smith (gps@star.sr.bham.ac.uk) or myself (rgm@roe.ac.uk) by Wednesday 6 May; offers from Early Career Researchers would be particularly welcome.

imSim is used by the Dark Energy Science Collaboration (DESC) to simulate Rubin images as seen by LSSTCam and ComCam. Working detector-by-detector, it reads a catalogue of objects (assume for now to be stars and galaxies) and determines which should be drawn for the detector at hand. To make sure it captures objects on the periphery of each image, it selects those lying inside its bounds plus a small 100-pixel buffer region.

This generally works well when considering the images from different detectors in isolation, but is limiting in a few circumstances. Diffraction spikes from bright stars can extend further than the buffer region. We’d also like to be able to simulate optical effects such as ghosting from internal reflections, where photons could end up anywhere at all. As it is now, photons moving off the edge of a detector are lost. To make sure you capture everything, you’d have to expand the buffer to the full size of the aperture; every detector would have to simulate every object and every photon, even though they’d ultimately only draw ~1/189 of the total. Some teams have had to go down this route, for example when simulating the collimated beam projector (CBP), but this is obviously extremely inefficient.

@William Lucas

What’s needed is a way to move photons between the processes handling different detectors MPI – this may sound ideal, except that we need to allow different detectors to run in different jobs, maybe at different times, potentially even on different systems. So, we use a simple form of message passing with files in a two-pass drawing process. At a high level, this takes place as follows:

1. In the first pass, simulate images for each detector as we always would, but we also have imSim collect any photons which move beyond the detector bounds and write them to file. We call those off-detector photons.

2. In the second pass, we generate a new set of images. These read as input the off-detector photons, and we then accumulate the flux from those photons on top of the images which were generated in the first pass.

This has required a few changes in imSim’s internal workings, particularly in order to prevent the multiple drawing of objects. With this, imSim can now simulate effects which cross detector boundaries without wasting compute resources, and the way is opened to more complex optics simulations.

@Hiranya Peiris is one of the latest scientists to be interviewed by Jim Al-Khalili on the Radio 4 science programme The Life Scientific.

Hiranya holds the Professorship of Astrophysics (1909) at the Institute of Astronomy, University of Cambridge. She is a Professorial Fellow at Murray Edwards College and a member of the Kavli Institute for Cosmology, Cambridge. 

In the interview, Hiranya talks about her early ambition to become an astronaut and how she was inspired to swap her studies from computer science to physics. She also explains how her work on the Legacy Survey of Space and Time will provide new perspectives on dark matter.

@Eleanor O'Kane

Useful resources

Share your passion

A reminder that you can access communications training videos, key messages and other communications resources for your LSST:UK presentations on the dedicated wiki page.

We’re looking for more people to take part in our In the Spotlight series. It’s a chance to tell us about your work and how you arrived at LSST:UK. If you’re interested, get in touch or fill in the webform.

I

@Eleanor O'Kane