用机器学习解析星系气体分布，揭示反馈机制如何影响星系晕内气体密度。
Interpretable machine learning of halo gas density profiles: a sensitivity analysis of cosmological hydrodynamical simulations
- 基于随机森林模型，从星系属性预测晕内气体密度分布。
- 总晕质量和中心星系气体质量是决定气体分布的最关键因素。
- 首次在全尺度模拟中量化反馈对气体分布的影响，适合宇宙学研究者。
恒星与活动星系核驱动的反馈过程影响气体在多尺度上的分布，从星系内部延伸至星际介质。然而，反馈如何通过关键星系属性塑造宿主晕的径向气体密度分布仍不明确。本文利用EAGLE、IllustrisTNG和Simba等宇宙学流体动力学模拟，开发了一种随机森林算法，基于晕总质量及中心星系的五项全局属性（气体与恒星质量、恒星形成率、黑洞质量与吸积率）预测晕内气体密度分布。该算法在 $10^{9.5} \ ext{M}_igodot < M_{\rm 200c} < 10^{15} \text{M}_\bigodot$ 晕质量范围与 $0<z<4$ 红移区间内，平均预测准确率达83-90%。首次对完整宇宙学模拟应用Sobol敏感性分析，量化各特征在不同距离上对气体密度的影响。所有模拟与红移下，晕总质量与中心星系气体质量最具影响力，而恒星与黑洞属性影响较小，其相对重要性随反馈模型与红移变化。该框架可嵌入半解析星系形成模型，实现与不同流体模拟一致的晕气体密度分布。同时为未来观测约束反馈模型提供概念验证。
原文摘要 · Abstract (English)
Stellar and AGN-driven feedback processes affect the distribution of gas on a wide range of scales, from within galaxies well into the intergalactic medium. Yet, it remains unclear how feedback, through its connection to key galaxy properties, shapes the radial gas density profile in the host halo. We tackle this question using suites of the EAGLE, IllustrisTNG, and Simba cosmological hydrodynamical simulations, which span a variety of feedback models. We develop a random forest algorithm that predicts the radial gas density profile within haloes from the total halo mass and five global properties of the central galaxy: gas and stellar mass; star formation rate; mass and accretion rate of the central black hole (BH). The algorithm reproduces the simulated gas density profiles with an average accuracy of $\sim$83-90% over the halo mass range $10^{9.5} \, \mathrm{M}_{\odot} < M_{\rm 200c} < 10^{15} \, \mathrm{M}_{\odot}$ and redshift interval $0<z<4$. For the first time, we apply Sobol statistical sensitivity analysis to full cosmological hydrodynamical simulations, quantifying how each feature affects the gas density as a function of distance from the halo centre. Across all simulations and redshifts, the total halo mass and the gas mass of the central galaxy are the most strongly tied to the halo gas distribution, while stellar and BH properties are generally less informative. The exact relative importance of the different features depends on the feedback scenario and redshift. Our framework can be readily embedded in semi-analytic models of galaxy formation to incorporate halo gas density profiles consistent with different hydrodynamical simulations. Our work also provides a proof of concept for constraining feedback models with future observations of galaxy properties and of the surrounding gas distribution.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。