发表机构
Lorentz Institute for Theoretical Physics, Leiden University; Leiden Observatory, Leiden University; Astrophysics Research Institute, Liverpool John Moores University; Institute for Computational Cosmology, Department of Physics, University of Durham（莱顿大学洛伦兹理论物理研究所; 莱顿大学莱顿天文台; 利物浦约翰摩尔斯大学天体物理学研究所; 杜伦大学物理学院计算宇宙学研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究利用贝叶斯推断方法，结合弱引力透镜峰值的红移、高度、角成团统计，可更精准约束宇宙学参数，其性能优于传统剪切两点相关函数
AI 中文摘要
弱引力透镜（WL）峰值可探测大尺度物质分布中两点透镜统计无法捕捉的非高斯信息。本研究采用模拟欧几里得（Euclid）分析的贝叶斯推断方法，探究高值WL峰值的高度分布、红移分布与角成团的宇宙学潜力。研究使用即将面世的暗物质仅含超立方体，其在十维空间中变化宇宙学参数，包括演化暗能量、中微子质量、衰变暗物质及标量谱指数的跑动。研究发现，各类WL峰值统计具有互补性：红移分布最适合约束物质密度Ωₘ与暗能量状态方程参数w₀、wₐ；高度分布与角成团对原初功率谱振幅ln(10¹⁰Aₛ)最敏感；三者结合还可探测重子密度Ωᵦ与哈勃参数h。与剪切两点相关函数对比显示，仅WL峰值就优于常用统计，三者结合约束更紧。在二维Ωₘ-ln(10¹⁰Aₛ)与w₀-wₐ参数平面，红移分布优于角成团与峰值高度分布。研究还分析了平滑尺度的影响，发现通常最小尺度效果最佳，多尺度结合时品质因数提升约2倍
英文摘要
Weak gravitational lensing (WL) peaks probe non-Gaussian information of the large-scale distribution of matter that is not captured by two-point lensing statistics. We study the cosmological potential of the height distribution, redshift distribution, and angular clustering of high-valued WL peaks using a Bayesian inference approach that mimics a $\textit{Euclid}$ analysis. We use a forthcoming dark-matter-only hypercube, which varies cosmology in a ten-dimensional space, including evolving dark energy, neutrino mass, decaying dark matter, and the running of the scalar spectral index. We find the individual WL peak statistics to be complementary, as the redshift distribution best constrains the matter density, $Ω_\mathrm{m}$, and the dark energy equation-of-state parameters, $w_0$, and $w_a$; the height distribution and angular clustering are most sensitive to the amplitude of the primordial power spectrum, $\ln(10^{10}A_\mathrm{s})$; while combining the three statistics allows us to also probe the baryon density, $Ω_\mathrm{b}$, and the Hubble parameter, $h$. A comparison to the shear two-point correlation function demonstrates that WL peaks alone outperform the commonly used statistics, while even tighter constraints are obtained when combining all. Considering the 2-dimensional $Ω_\mathrm{m}-\ln(10^{10}A_\mathrm{s})$ and $w_0-w_a$ parameter planes, we find that the redshift distribution outperforms the angular clustering and peak-height distributions. We study the impact of the smoothing scale and find that, typically, the smallest scales yield the best results and the figure of merit improves by a factor $\approx2$ when combining multiple scales.
Comments20 pages, 13 figures (including appendices), accepted for publication in MNRAS
