Search for: All records

ABSTRACT The Eos cloud, recently discovered in the far ultraviolet via H$$_2$$ fluorescence, is one of the nearest known dark molecular clouds to the Sun, with a distance spanning from $${\sim} 94\rm{\!-\!}136$$ pc. However, with a mass ($${\sim} 5.5\times 10^3$$ M$$_\odot$$) just under $$40\,$$ per cent that of star forming clouds like Taurus and evidence for net molecular dissociation, its evolutionary and star forming status is uncertain. We use Gaia data to investigate whether there is evidence for a young stellar population that may have formed from the Eos cloud. Comparing isochrones and pre-main sequence evolutionary models there is no clear young stellar population in the region. While there are a small number of $${<} 10$$ Myr stars, that population is statistically indistinguishable from those in similar search volumes at other Galactic latitudes. We also find no unusual spatial or kinematic clustering toward the Eos cloud over distances $$70\!-\!150$$ pc. Overall, we conclude that the Eos cloud has most likely not undergone any recent substantial star formation and further study of the dynamics of the cloud is required to determine whether it will do so in the future. 

Abstract Advancing our understanding of astrophysical turbulence is bottlenecked by the limited resolution of numerical simulations that may not fully sample scales in the inertial range. Machine-learning (ML) techniques have demonstrated promise in upscaling resolution in both image analysis and numerical simulations (i.e., superresolution). Here we employ and further develop a physics-constrained convolutional neural network ML model called “MeshFreeFlowNet” (MFFN) for superresolution studies of turbulent systems. The model is trained on both the simulation images and the evaluated partial differential equations (PDEs), making it sensitive to the underlying physics of a particular fluid system. We develop a framework for 2D turbulent Rayleigh–Bénard convection generated with theDedaluscode by modifying the MFFN architecture to include the full set of simulation PDEs and the boundary conditions. Our training set includes fully developed turbulence sampling Rayleigh numbers (Ra) ofRa= 10⁶–10¹⁰. We evaluate the success of the learned simulations by comparing the power spectra of the directDedalussimulation to the predicted model output and compare both ground-truth and predicted power spectral inertial range scalings to theoretical predictions. We find that the updated network performs well at allRastudied here in recovering large-scale information, including the inertial range slopes. The superresolution prediction is overly dissipative at smaller scales than that of the inertial range in all cases, but the smaller scales are better recovered in more turbulent than laminar regimes. This is likely because more turbulent systems have a rich variety of structures at many length scales compared to laminar flows.