Periodic Reporting for period 1 - PiCOGAMBAS (Precision Cosmology with Galaxy and Microwave Background surveys)
Période du rapport: 2021-02-01 au 2023-01-31
Understanding the universe we live in is one of the challenges that pushed mankind to look at the sky since the dawn of civilization. In modern times, accurate astrophysical observations have allowed us to turn questions like "how did the universe begin?" or "what the universe if made of?" into scientific problems.
The standard cosmological model predicts that the universe started in a hot and dense phase commonly known as Big Bang and then underwent an accelerated expansion called inflation where the seeds of the matter density perturbation that later evolved in galaxies and stars were generated. After inflation the universe expanded slowly since the beginning of a new accelerated phase a few billion years ago driven by an unknown energy component called dark energy. Understanding the nature of inflation and dark energy are the two of the main questions that cosmologists hope to answer in the next decade using observations of the large scale structures (LSS) of the universe (galaxies, galaxy clusters etc.) and of the Cosmic Microwave Background (CMB), the relic light of the Big Bang.
During its journey towards us, the CMB interacts with the LSS which leave an imprint in the properties of CMB photons. Combining observations of the CMB and data of galaxy surveys observing the LSS we can thus study the properties of the matter distribution of the universe and its evolution with two different and complementary techniques, correlating mass maps acquired with different techniques (cross-correlation). This will sharpen our understanding of the universe beyond what can be normally achieved and beyond the limiting factor affecting the analysis of both kinds of data sets when performed separately.
With the new generation of CMB experiments (such as Simons Observatory) and galaxy surveys (such as the ESA mission Euclid) all observing at the same time the same sky by the end of 2023, this field has the potential to deliver multiple "firsts" in cosmology in the coming years.
To fully exploit this potential we need to better understand the statistical methods involved in the analysis of the data, test their robustness, design new techniques and observables capable of extracting the maximum amount of information from the data. All these questions are at the core of the PiCOGAMBAS project that tries to tackle these challenges using theoretical modeling and numerical techniques.
The standard cosmological model predicts that the universe started in a hot and dense phase commonly known as Big Bang and then underwent an accelerated expansion called inflation where the seeds of the matter density perturbation that later evolved in galaxies and stars were generated. After inflation the universe expanded slowly since the beginning of a new accelerated phase a few billion years ago driven by an unknown energy component called dark energy. Understanding the nature of inflation and dark energy are the two of the main questions that cosmologists hope to answer in the next decade using observations of the large scale structures (LSS) of the universe (galaxies, galaxy clusters etc.) and of the Cosmic Microwave Background (CMB), the relic light of the Big Bang.
During its journey towards us, the CMB interacts with the LSS which leave an imprint in the properties of CMB photons. Combining observations of the CMB and data of galaxy surveys observing the LSS we can thus study the properties of the matter distribution of the universe and its evolution with two different and complementary techniques, correlating mass maps acquired with different techniques (cross-correlation). This will sharpen our understanding of the universe beyond what can be normally achieved and beyond the limiting factor affecting the analysis of both kinds of data sets when performed separately.
With the new generation of CMB experiments (such as Simons Observatory) and galaxy surveys (such as the ESA mission Euclid) all observing at the same time the same sky by the end of 2023, this field has the potential to deliver multiple "firsts" in cosmology in the coming years.
To fully exploit this potential we need to better understand the statistical methods involved in the analysis of the data, test their robustness, design new techniques and observables capable of extracting the maximum amount of information from the data. All these questions are at the core of the PiCOGAMBAS project that tries to tackle these challenges using theoretical modeling and numerical techniques.
One of the main objectives of the project consists in producing robust measurements (lensing reconstruction) of the total matter distribution in the universe (CMB lensing potential) from observation of the CMB in temperature and polarization. CMB lensing is the key observable for cross-correlation with LSS and to probe the nature of neutrinos. Using realistic simulations of the data acquisition chains of a CMB experiment I demonstrated that lensing reconstruction will be unaffected by instrumental systematicatics introduced by the next generation instruments and provided a public tool to study further such problems (Fabbian & Peloton 21, Mirmelstein, Fabbian+21). I also discovered a new effect in the lensing reconstruction measurements appearing when we remove regions with bright astrophysical objects from the analysis. Since these objects live in the same matter clumps that lens the CMB, removing them will introduce errors in the lensing reconstruction that can be severe if unaccounted and particularly important for cross-correlations with LSS tracers.
Alongside these works, I studied new applications of cross-correlation techniques. Together with colleagues of the Euclid consortium, I forecasted for the first time the performances of cross-correlation with Euclid (galaxy clustering, galaxy lensing) and CMB lensing in constraining cosmological parameters (Euclid collaboration+22). We showed that such techniques will improve by more than 2x or more the errors on dark energy models achievable using Euclid data alone, thus providing new analysis possibilities.
I also worked on the first public statistical tool to exploit the decay of gravitational potentials induced by dark energy (ISW effect) to constrain models of dark energy and modified gravity from CMB and CMB lensing measurements of the Planck satellite (Carron, Lewis, Fabbian+22) at large scales, which will be the reference for the coming decade.
I studied how future CMB experiments such as the SO can also produce transformational measurements in astrophysics. In particular I demonstrated that SO will be able to map with high sensitivity and for the first time the polarization of the CO emission lines in molecular clouds over the full sky, allowing us to learn the role of the magnetic field in the process of star formation (Hensley+2022).
The last main topic I investigated is constraining the physics of the early universe using CMB B-mode polarization (generated during inflation) together with galaxy surveys. With colleagues of the POLARBEAR collaboration I produced the 2nd best measurement of inflation B-modes from Earth (Takakura+22). The inflationary B-mode is in fact contaminated by spurious B-modes induced by gravitational lensing which need to be subtracted (delensing). Together with colleagues of SO we showed that using cross-correlations of galaxy clustering, CMB lensing and infrared galaxies we can improve the current measurements on inflation by more than 20x (Namikawa+22). Alternative probes of inflation consists in measuring deviations from the Gaussian statistics of the density perturbations. This can be achieved by studying CMB anisotropies or the galaxy distribution but new information can be extracted using local deviations from the CMB emission law (spectral distortions). In Bianchini & Fabbian 22 I reanalysed the last of these measurements (released in the late 90s) with modern statistical techniques and improved them by a factor of 2x. This allowed me to produce a map of those distortions and correlate it for the first time with CMB temperature and polarization maps to obtain constraints on primordial non-Gaussianities in a regime never investigated before, providing the first proof of concept for such techniques.
Alongside these works, I studied new applications of cross-correlation techniques. Together with colleagues of the Euclid consortium, I forecasted for the first time the performances of cross-correlation with Euclid (galaxy clustering, galaxy lensing) and CMB lensing in constraining cosmological parameters (Euclid collaboration+22). We showed that such techniques will improve by more than 2x or more the errors on dark energy models achievable using Euclid data alone, thus providing new analysis possibilities.
I also worked on the first public statistical tool to exploit the decay of gravitational potentials induced by dark energy (ISW effect) to constrain models of dark energy and modified gravity from CMB and CMB lensing measurements of the Planck satellite (Carron, Lewis, Fabbian+22) at large scales, which will be the reference for the coming decade.
I studied how future CMB experiments such as the SO can also produce transformational measurements in astrophysics. In particular I demonstrated that SO will be able to map with high sensitivity and for the first time the polarization of the CO emission lines in molecular clouds over the full sky, allowing us to learn the role of the magnetic field in the process of star formation (Hensley+2022).
The last main topic I investigated is constraining the physics of the early universe using CMB B-mode polarization (generated during inflation) together with galaxy surveys. With colleagues of the POLARBEAR collaboration I produced the 2nd best measurement of inflation B-modes from Earth (Takakura+22). The inflationary B-mode is in fact contaminated by spurious B-modes induced by gravitational lensing which need to be subtracted (delensing). Together with colleagues of SO we showed that using cross-correlations of galaxy clustering, CMB lensing and infrared galaxies we can improve the current measurements on inflation by more than 20x (Namikawa+22). Alternative probes of inflation consists in measuring deviations from the Gaussian statistics of the density perturbations. This can be achieved by studying CMB anisotropies or the galaxy distribution but new information can be extracted using local deviations from the CMB emission law (spectral distortions). In Bianchini & Fabbian 22 I reanalysed the last of these measurements (released in the late 90s) with modern statistical techniques and improved them by a factor of 2x. This allowed me to produce a map of those distortions and correlate it for the first time with CMB temperature and polarization maps to obtain constraints on primordial non-Gaussianities in a regime never investigated before, providing the first proof of concept for such techniques.
The project has so far been on track with its goals and schedule and produced important results in the field. Several analyses are however still ongoing, including:
- The first analysis on cosmological simulations of the impact of massive neutrinos on cosmic voids ( the most underdense regions in the universe) as identified in Euclid and CMB lensing data.
- New constraints on the amplitude of density fluctuations and primordial non-Gaussianity from a new catalog of distant quasars (a type of luminous galaxies powered by supermassive black holes)
- The first comprehensive study of the properties of infrared galaxies a function of models of the physics of gas.
- The first assessment of the impact of non-linear effects in a new, optimal class of statistical estimator of lensing reconstruction and delensing.
- The first constraints on model of galaxy formation achieved with machine learning and measurements of CMB spectral distortions.
These projects will potentially deliver some of the best cosmological constraints in the field of cross-correlation and will deliver new robust methods to analyze the data of next generation experiments.
- The first analysis on cosmological simulations of the impact of massive neutrinos on cosmic voids ( the most underdense regions in the universe) as identified in Euclid and CMB lensing data.
- New constraints on the amplitude of density fluctuations and primordial non-Gaussianity from a new catalog of distant quasars (a type of luminous galaxies powered by supermassive black holes)
- The first comprehensive study of the properties of infrared galaxies a function of models of the physics of gas.
- The first assessment of the impact of non-linear effects in a new, optimal class of statistical estimator of lensing reconstruction and delensing.
- The first constraints on model of galaxy formation achieved with machine learning and measurements of CMB spectral distortions.
These projects will potentially deliver some of the best cosmological constraints in the field of cross-correlation and will deliver new robust methods to analyze the data of next generation experiments.