AI 中文总结
本研究将四种动力学方法应用于玉夫座星表，测得其内暗物质斜率Γ=2.49⁺⁰·²³₋₀·₂₀，发现偏差主要源于几何结构，且其局部斜率与NFW轮廓一致，为核心-尖峰问题提供了新的观测约束。
AI 中文摘要
核心-尖峰问题是暗物质密度轮廓内斜率的模拟结果与观测结果之间的差异。双总体估计器将矮椭球星系按金属丰度拆分，是打破质量-各向异性简并的标准方法。本研究将四种动力学方法应用于1339颗恒星的Tolstoy等人（2023）VLT/FLAMES玉夫座星表，这四种方法分别是：球形Jeans法、GravSphere法、Walker和Penarrubia（2011）双总体估计器，以及基于连续作用量的分布函数f(J,[Fe/H])。本研究得到Γ=2.49⁺⁰·²³₋₀·₂₀，其中99.6%的后验分布位于Γ=2之上（经λ校准后为84.4%），该结果对成员星筛选和贫金属污染具有鲁棒性；GravSphere法得到γ=0.46⁺⁰·⁴⁰₋₀·₂₈。通过控制示踪粒子几何结构固定、打破金属丰度-运动学关联的控制模拟，发现77%-83%的恢复偏差源于几何结构而非拆分本身；因此，基于模拟的总体拆分研究必须控制几何结构。连续分布函数编码了一种任何基于拆分的方法都无法获取的金属丰度-作用量耦合k_J；对575颗具有Gaia自行的恒星进行直接作用量空间计算，发现金属丰度[Fe/H]在径向作用量的三个数量级上下降了0.3 dex。在Read等人（2019）的中心密度诊断上，相同的后验分布给出ρ_DM(150 pc)=1.77×10⁸ M⊙ kpc⁻³，与他们独立的玉夫座测量值相差0.74σ，将玉夫座归入其尖峰类；本研究的浅γ是渐近量，150 pc处的局部斜率与Navarro-Frenk-White轮廓一致。因此，与反馈驱动的核心形成模型的任何张力仅取决于斜率轴。
英文摘要
The core-cusp problem is a discrepancy between simulations and observations of the slopes of inner dark matter density profiles. Splitting dwarf spheroidals by metallicity is a standard method to break the mass-anisotropy degeneracy. Four dynamical methods are applied to the 1339-star Tolstoy et al. (2023) VLT/FLAMES Sculptor catalog: spherical Jeans, GravSphere, the Walker and Peñarrubia (2011) two-population estimator, and a continuous $f(\boldsymbol{J},[\text{Fe}/\text{H}])$ action-based distribution function. This work finds $Γ=2.49^{+0.23}_{-0.20}$, with 99.6\% of the posterior lying above $Γ=2$ (84.4\% after $λ$ calibration), robust to membership cuts and metal-poor contamination; GravSphere gives $γ=0.46^{+0.40}_{-0.28}$. Controlled mocks that hold tracer geometry fixed while breaking the metallicity-kinematics link attribute $\sim77$--$83\%$ of the recovered bias to geometry rather than decomposition itself. The continuous distribution function encodes a metallicity--action coupling $k_J$ inaccessible to any split-based method; a direct action-space computation on 575 stars with Gaia proper motions finds $[\text{Fe}/\text{H}]$ declining by $\sim0.3\,\text{dex}$ across three decades in radial action. On the central-density diagnostic of Read et al. (2019), the same posteriors give $ρ_\text{DM}(150\,\mathrm{pc})=1.77\times10^8\,M_\odot\,\mathrm{kpc^{-3}}$, agreeing with their independent Sculptor measurement at $0.74σ$ and placing Sculptor in their cusp-like class. The shallow $γ$ of this work is an asymptotic quantity; the local slope at $150\,\mathrm{pc}$ is consistent with the Navarro--Frenk--White profile under the spherical models used here, though a recent axisymmetric analysis using resolved 3D velocities finds a shallow slope over that range and identifies an inclination degeneracy to which spherical models are blind.
Comments30 pages, 22 figures. Code and MCMC chains: https://doi.org/10.5281/zenodo.22052154
