Periodic Reporting for period 1 - OCULIS (Observational Cosmology Using Large Imaging Surveys)
Periodo di rendicontazione: 2023-09-01 al 2026-02-28
The current standard model of cosmology successfully describes a wide variety of measurements, but its main ingredients, dark matter and dark energy, are a great mystery. The observations demonstrate that our theories of particle physics and gravity are either incomplete or incorrect, but we lack compelling theoretical guidance to solve this crisis in cosmology. The only viable course of action is to improve observational constraints by at least an order of magnitude. This is the objective of Euclid, the recently launched ESA mission that will survey more than half of the extragalactic sky over a period of six year. Its homogeneous, high-quality space data are unaffected by atmospheric blurring and infrared emission, resulting in unrivaled data fidelity, but the cosmological interpretation of the Euclid data is complex as information from multiple probes needs to be combined, while the observational signatures of new physics are subtle. Incorrect modelling of the galaxy populations, astrophysical processes or instrumental effects can easily be mistaken as evidence for new physics. Although Euclid is designed to minimize observational biases, these cannot be fully eliminated, and a thorough understanding of the intricacies of the data is essential. Similarly, the main probes were chosen to be robust, but they are not immune to astrophysics. To fully exploit the unprecedented precision of Euclid we need to disentangle the physics of galaxy formation and cosmology.
OCULIS focuses on several key areas where progress is needed to realize the full potential of the Euclid data. First, we will improve the actual measurement of the lensing signal by using in-flight data to correct for instrumental effects that would otherwise bias the shape estimates. OCULIS will also exploit the high-quality data from Euclid to improve our understanding of astrophysical sources of bias, in particular the effect of intrinsic alignments of galaxies that contaminate the lensing signal. By developing dedicated measurement tools we will explore the alignment signal as a function of galaxy properties, and compare these with predictions from state-of-the-art cosmological hydrodynamic simulations.
Thanks to the superb image quality of Euclid, the lensing signal can be measured on angular scales that are inaccessible with ground-based data. In particular, the lensing signal at small radii around foreground lenses can be used to determine the stellar and halo masses with a minimal reliance on the cosmological model and to study the tidal stripping of dark matter halos as a function of environment. This requires modifications to the standard shape measurement pipelines, so that they can account for the contaminating light from the lens galaxies. The improved shape measurement algorithms will also used to measure the intrinsic alignments of galaxies that contaminate the lensing signal. By relating the resulting measurements to the surrounding matter distribution we can advance a physical understanding of these phenomena, and use the findings to improve the fidelity of the modelling of small-scale astrophysics in hydrodynamical simulations. To predict the various cosmological signals consistently, whilst accounting for the uncertainty in the astrophysics (and fundamental physics), we will incorporate our results into an emulator, which enable the best possible measurements of cosmological parameters using Euclid.
In short, the aim of OCULIS is to use the Euclid data themselves to better understand the observational sources of bias, and using these findings to improve the overall interpretation of the cosmological signal. This allows us to probe smaller scales and use fainter galaxies, thus improving the overall ability of the mission to determine cosmological parameters.
OCULIS focuses on several key areas where progress is needed to realize the full potential of the Euclid data. First, we will improve the actual measurement of the lensing signal by using in-flight data to correct for instrumental effects that would otherwise bias the shape estimates. OCULIS will also exploit the high-quality data from Euclid to improve our understanding of astrophysical sources of bias, in particular the effect of intrinsic alignments of galaxies that contaminate the lensing signal. By developing dedicated measurement tools we will explore the alignment signal as a function of galaxy properties, and compare these with predictions from state-of-the-art cosmological hydrodynamic simulations.
Thanks to the superb image quality of Euclid, the lensing signal can be measured on angular scales that are inaccessible with ground-based data. In particular, the lensing signal at small radii around foreground lenses can be used to determine the stellar and halo masses with a minimal reliance on the cosmological model and to study the tidal stripping of dark matter halos as a function of environment. This requires modifications to the standard shape measurement pipelines, so that they can account for the contaminating light from the lens galaxies. The improved shape measurement algorithms will also used to measure the intrinsic alignments of galaxies that contaminate the lensing signal. By relating the resulting measurements to the surrounding matter distribution we can advance a physical understanding of these phenomena, and use the findings to improve the fidelity of the modelling of small-scale astrophysics in hydrodynamical simulations. To predict the various cosmological signals consistently, whilst accounting for the uncertainty in the astrophysics (and fundamental physics), we will incorporate our results into an emulator, which enable the best possible measurements of cosmological parameters using Euclid.
In short, the aim of OCULIS is to use the Euclid data themselves to better understand the observational sources of bias, and using these findings to improve the overall interpretation of the cosmological signal. This allows us to probe smaller scales and use fainter galaxies, thus improving the overall ability of the mission to determine cosmological parameters.
Euclid was successfully launched on July 1st 2023. Following the commissioning, the PI was deeply involved in the performance verification activities. These initial data allowed us to finetune algorithms to process the data, which is one of the objectives of OCULIS. One surprise was that the regular approach to detect and account for cosmic rays did not perform as well as planned. Members of the OCULIS team carried out work to quantify this and explored the performance of a deep learning algorithm, which is now being implemented into the pipeline.
Team members also plays a prominent role in the calibration of the shear measurement algorithms, which relies on detailed image simulations. To reduce the reliance of such simulated data, we have demonstrated the Metacalibration is particularly powerful. This approach is now part of the main pipeline, while we have also applied it to the KiloDegree Survey. As these calibrations remain work in progress, the PI has explored a new mitigation strategy that appears to work well.
Accurate knowledge of the blurring of the galaxy images by the telescope optics is essential for the success of Euclid. We have explored the possibility of using diffraction spikes observed around bright stars to monitor wavefront errors. This novel approach
works remarkably well, showing that the current model is incomplete. The approach is extremely powerful, with potential application for future space-based missions, while the algorithm has been implemented in the Euclid pipeline.
We have explored how we can improve the information on IA that can be extracted from cosmological hydrodynamic simulations. Using data from the PAU Survey, we were also able to extend IA measurements to fainter galaxies. These insights are used to explore the best way to model the IA contamination for the first Euclid data release (DR1). One of the team members is leading the paper describing the findings. Work to derive a more physically motivated model has started, in particular aiming to capture the role of mergers. To this end, we will examine the IA signal using a new suite of hydrodynamic simulations developed by colleagues in Leiden.
Another important pillar of OCULIS is to improve the connection between the observed properties of galaxies and their surrounding dark matter distribution. We have demonstrated that this is a worthwhile objective to pursue. Work on simple simulated data show that we should be able to constrain the stellar masses of galaxies directly, thus further constraining the stellar-mass-halo-mass relation. This is an important input to improve the calibration of the sub-grid physics in hydrodynamic simulations. We also started exploring if the stellar-mass-halo-mass relation depends on environment using both simulations and KiDS data. The challenge is the modelling, because the underlying halo mass function depends on the environment. This is an important first step to advancing the modelling of the signal in Euclid data.
Team members also plays a prominent role in the calibration of the shear measurement algorithms, which relies on detailed image simulations. To reduce the reliance of such simulated data, we have demonstrated the Metacalibration is particularly powerful. This approach is now part of the main pipeline, while we have also applied it to the KiloDegree Survey. As these calibrations remain work in progress, the PI has explored a new mitigation strategy that appears to work well.
Accurate knowledge of the blurring of the galaxy images by the telescope optics is essential for the success of Euclid. We have explored the possibility of using diffraction spikes observed around bright stars to monitor wavefront errors. This novel approach
works remarkably well, showing that the current model is incomplete. The approach is extremely powerful, with potential application for future space-based missions, while the algorithm has been implemented in the Euclid pipeline.
We have explored how we can improve the information on IA that can be extracted from cosmological hydrodynamic simulations. Using data from the PAU Survey, we were also able to extend IA measurements to fainter galaxies. These insights are used to explore the best way to model the IA contamination for the first Euclid data release (DR1). One of the team members is leading the paper describing the findings. Work to derive a more physically motivated model has started, in particular aiming to capture the role of mergers. To this end, we will examine the IA signal using a new suite of hydrodynamic simulations developed by colleagues in Leiden.
Another important pillar of OCULIS is to improve the connection between the observed properties of galaxies and their surrounding dark matter distribution. We have demonstrated that this is a worthwhile objective to pursue. Work on simple simulated data show that we should be able to constrain the stellar masses of galaxies directly, thus further constraining the stellar-mass-halo-mass relation. This is an important input to improve the calibration of the sub-grid physics in hydrodynamic simulations. We also started exploring if the stellar-mass-halo-mass relation depends on environment using both simulations and KiDS data. The challenge is the modelling, because the underlying halo mass function depends on the environment. This is an important first step to advancing the modelling of the signal in Euclid data.
As the start of the survey was delayed, we have had limited opportunity to publish our key scientific results. However, the most surprising finding to date is the potential of using diffraction spikes to determine the focus of an optical system. This technique is not widely known, and the work by PhD student Dennis Neumann has shown this is powerful in the case of Euclid. Although current work focuses on monitoring the instrument performance, it is possible that this technique can be applied to other telescopes (such as the Roman Space Telescope) and possible other optical systems.