论文标题
部分可观测时空混沌系统的无模型预测
The Young and the Wild: What happens to Protoclusters forming at z = 4?
论文作者
论文摘要
储层计算是预测湍流的有力工具,其简单的架构具有处理大型系统的计算效率。然而,其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间,然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用,该系统模拟了湍流的若干特征。我们表明,使用径向基函数作为非线性投影器,即使只有部分观测并且不知道控制方程,也能稳健地捕捉复杂的系统非线性。最后,我们表明,当测量稀疏、不完整且带有噪声,甚至控制方程变得不准确时,我们的网络仍然可以产生相当准确的预测,从而为实际湍流系统的无模型预测铺平了道路。
Using one of the largest volumes of the hydrodynamical cosmological simulation suit Magneticum, we study the evolution of protoclusters identified at redshift = 4, with properties similar to SPT2349-56. We identify 42 protoclusters in the simulation, as massive and equally rich in substructures as observed, confirming that these structures are already virialized. The dynamics of the internally fast rotating member galaxies within these protoclusters resembles observations, merging rapidly to form the cores of the BCGs of the assembling clusters. Half of the gas reservoir of these structures is in a hot phase, with the metal-enrichment at a very early stage. These systems show a good agreement with the observed amount of cold star-forming gas, largely enriched to solar values. We predict that some of the member galaxies are already quenched at z = 4, rendering them undetectable through measurements of their gas reservoir. Tracing the evolution of protoclusters reveals that none of the typical mass indicators at high redshift are good tracers to predict the present-day mass of the system. We find that none of the simulated protoclusters with properties as SPT2349-56 at z = 4.3, are among the top ten most massive clusters at redshift z = 0, with some barely reaching masses of M = 2 x 10^14Msun. Although the average star-formation and mass-growth rates in the simulated galaxies match observations at high redshift reasonably well, the simulation fails to reproduce the extremely high total star-formation rates within observed protoclusters, indicating that the sub-grid models are lacking the ability to reproduce higher star-formation efficiency (or lower depletion timescales).