Sub-resolution modelling of astrophysical processes

在宇宙学流体动力学模拟中,涉及星系形成和演化的天体物理过程跨越了巨大的动力学范围。它们从约$\sim$Mpc尺度——引力不稳定性驱动暗物质组分演化与暗晕等级式并合之处——一直延伸到约$\sim$pc和$\sim$亚pc尺度——黑洞吸积和恒星形成等过程发生之处,途经约$\sim$kpc尺度——例如星系风将恒星反馈能量分配到周围介质之处。 因此,对次网格物理的需求至关重要:次分辨率模型描述的是发生在宇宙学流体动力学模拟分辨率极限以下的过程,但这些过程会影响在显式解析尺度上演化的模拟结构。次分辨率方案通常借助相当简单的解析或理论模型,辅以恰当选择并经校准以重现观测的参数,对相当复杂的过程进行唯象描述(另见§3.16小节)。

The astrophysical processes involved in the formation and evolution of galaxies in cosmological hydrodynamical simulations span a huge dynamical range of scales. They indeed vary from the $\sim$Mpc scales where gravitational instabilities drive the evolution of the DM component and the hierarchical assembly of haloes, down to the $\sim$parsec and $\sim$sub-parsec scales, where processes like BH accretion and star formation take place, going through the $\sim$kpc scales, where e.g. galactic winds distribute the stellar feedback energy to the ambient medium. The call for sub-grid physics is thus essential: sub-resolution models account for processes that occur below the resolution limit of cosmological hydrodynamical simulations, but that affect the evolution of the simulated structure on scales that are explicitly resolved. Sub-resolution prescriptions usually resort to rather simple analytical or theoretical models, and/or to the phenomenological description of rather complex processes, through a suitable choice of parameters that are calibrated to reproduce observations (see also sub-section §3.16).

数值天体物理学常受到的一种批评,便是建模物理过程时所使用的参数数量过多,特别是次分辨率模型中的参数。然而,仅靠微调参数并无法重现观测或预期结果。结果更主要地取决于过程的恰当参数化方式。例如:在建模恒星形成时(见§3.12小节),结果主要由恒星形成率(SFR)与实际燃料(如可用的分子气体或冷气体质量)之间的关系所决定,而非上述两个量之间比例常数的精确取值。 因此,参数空间探索的目的并非为模拟所要比较的观测提供最佳拟合。它更旨在深化对模型依赖于不同未知量的理解,并对约束较弱的物理量做出预测。

A common criticism that has been often raised to numerical astrophysics deals with the number of parameters that are employed when modelling physical processes, in particular when referring to the parameters that enter sub-resolution models. However, observed or expected results cannot be reproduced by just fine-tuning parameters. Results are rather determined by proper parametrizations of processes. For instance: when modelling star formation (see sub-section §3.12), results are mainly driven by the way in which the star formation rate (SFR) relates to the actual fuel (e.g. the mass of molecular or cold gas available), rather than to the exact value of the proportionality constant between the two aforementioned quantities. Parameter space exploration is thus not meant to provide the best fit to observations that simulations want to compare with. It rather aims at enabling a better understanding of the dependence of a model on different unknowns and at making predictions for loosely constrained physical quantities.

在接下来的小节中,我们将回顾对星系形成和演化影响最为显著的天体物理过程。它们均自洽地包含在最先进的宇宙学流体动力学模拟中。

In the next sub-sections, we will review the most important astrophysical processes that significantly affect galaxy formation and evolution. They are all self-consistently included in state-of-the-art cosmological hydrodynamical simulations.


系列导航 ← 上一篇:§3.9 → 下一篇:§3.11