A procedure for modelling asymptotic g‐mode pulsators: The case of γ Doradus stars

Mode identification is one of the first and main problems we encounter in trying to develop the complete potential of asteroseismology. In the particular case of g‐mode pulsators, this is still an unsolved problem, from both the observational and theoretical points of view. Nevertheless, in recent years, some observational and theoretical efforts have been made to find a solution. In this work we use the latest theoretical and computational tools to understand asymptotic g‐mode pulsators: 1) the Frequency Ratio Method, and 2) Time Dependent Convection. With these tools, a self‐consistent procedure for mode identification and modelling of these g‐mode pulsators can be constructed. This procedure is illustrated using observational information available for the γ Doradus star 9Aurigae. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

一个大质量恒星形成复合体的动力学性质

恒星形成于分子云之中, 分子外向流是恒星形成正在进行的重要动力学特征, 也是研究和认识恒星形成的重要契入点. 利用紫金山天文台青海观测站德令哈13.7m毫米波望远镜, 采用5种分子谱线探针(包括¹²CO、¹³CO、C¹⁸O、HCO$^+$ $J=1-0$和CS $J=2-1$, J为角动量量子数), 对一个包含IRAS 19230+1506、IRAS 19232+1504和G050.3179--00.4186这3个源的大质量恒星形成复合体进行了成图观测研究. 通过对以上分子谱线数据并结合红外波段巡天数据的分析, 在这3个源中首次探测到了分子外向流活动, 并确定了分子外向流的中心驱动源. 最后对这3个源进行了分子外向流相关物理量参数的计算, 分析了这些物理量参数之间的关系, 结果表明分子外向流的性质与中心驱动源的性质息息相关.     

Yields of Nucleosynthesis from massive and intermediate mass stars and constraints on their final evolution

Nucleosynthetic yields and production rates of helium and heavy elements are derived using new initial mass functions which take into account the recent revisions in O star counts and the stellar models of Maeder (1981a, b) which incorporate the effects of massloss on evolution. The current production rates are significantly higher than the earlier results due to Chiosi & Caimmi (1979) and Chiosi (1979), and a near-uniform birthrate operating over the history of the galactic disc explains the currently observed abundances. However, the yields are incompatibly high, and to obtain agreement it is necessary to assume that stars above a certain mass do not explode but proceed to total collapse. Further confirmation of this idea comes from the consideration of the specific yields and production rates of oxygen, carbon and iron and the constraints imposed by the observational enrichment history in the disc as discussed by Twarog & Wheeler (1982). Substantial amounts of⁴He and¹⁴C, amongst the primary synthesis species, are contributed by the intermediate mass stars in their wind phases. If substantial numbers of them exploded as Type I SN, their contribution to the yields of¹²C and⁵⁶Fe would be far in excess of the requirements of galactic nucleosynthesis. Either efficient massloss precludes such catastrophic ends for these stars, or the current stellar models are sufficiently in error to leave room for substantial revisions in the specific yields. The proposed upward revision of the¹²C (α,γ)¹⁶O rate may produce the necessary changes in stellar yields to provide a solution to this problem. Stars that produce most of the metals in the Galaxy are the same ones that contribute most to the observed supernova rate.     

V2213 Cyg is a New W Uma-Type System

V2213 Cyg was discovered as a variable star by Pavlenko (1999) in 1998. We present our photometry of V2213 Cyg from 1998–2003 based on CCD observations with the K-380 Cassegrain telescope of CrAO and the 60 cm Zeiss telescope of SAI. Observations have been carried out mostly in R and sometimes in B and V Johnson system. The total amount of data is 2270 points, covering ∼50 nights. We classify this binary as a W UMa-type contact system. Using all data we determined the orbital period to be 0.350079 ± 0.000007 day. The mean brightness varies between R = 14.35 and 14.05. The mean 1999–2003 orbital light curve has two humps and a primary minimum (I), which is 0.04 mag brighter than the deeper secondary one (II). The mean humps have slightly different height. The difference between two individual maxima varies within 0.1 mag, which may indicate an activity of the components. The highest hump is an asymmetrical one: it has sort of a shoulder at phases 0.75–0.80, before entering the less deep primary minimum (phase 0.0). The system is rather reddened, its colour indices are: B−V ∼ 0.8 and V−R ∼ 0.7, and give a spectral class of V2213 Cyg earlier than K.     

Sputtering in the outflows of cool stars

Radiation pressure acts to accelerate dust grains and, by transfer of momentum through collisions with the gas, drives the outflows of late-type stars. Some of these dust–gas collisions may be energetic enough to remove atoms from the dust grains. From an assumed initial size distribution for the dust grains, the method of Krüger et al. is used to study the evolution of a sample of spherical amorphous carbon grains under conditions typical of a late-type star. The size distribution of dust grains is presented for various sets of model parameters. One set of models assumes an initial Mathis, Rumpl & Nordsieck (MRN) distribution for the dust grains. The high-luminosity ( L ∗), high-effective temperature ( T _eff) set of parameters has a terminal velocity ( v _term) that is near, but above , the upper limit of observed outflow velocities for carbon stars (∼30 km s⁻¹ for the assumed ̇ of 5×10⁻⁶ M_⊙ yr⁻¹). The low L ∗, T _eff model has a v _term that lies near, but below , the upper limit of observed velocities. A significant amount of sputtering occurs in the high L ∗, T _eff model with ∼40 per cent of the grain mass sputtered. About ∼1 per cent of the dust mass is sputtered in the low L ∗, T _eff. Another set of models assumes that the dust forms with a log-normal distribution. Here, v _term is nearly the same for the high L ∗, T _eff model as for the low L ∗, T _eff model. This is a result of the large amount of dust mass loss (∼75 per cent) by sputtering in the high L ∗, T _eff model.     

基于Stacking集成学习的恒星/星系分类研究

机器学习在当今诸多领域已经取得了巨大的成功,但是机器学习的预测效果往往依赖于具体问题.集成学习通过综合多个基分类器来预测结果,因此,其适应各种场景的能力较强,分类准确率较高.基于斯隆数字巡天(Sloan Digital Sky Survey,SDSS)计划恒星/星系中最暗源星等集分类正确率低的问题,提出一种基于Stacking集成学习的恒星/星系分类算法.从SDSS-DR7(SDSS Data Release 7)中获取完整的测光数据集,并根据星等值划分为亮源星等集、暗源星等集和最暗源星等集.仅针对分类较为复杂且困难的最暗源星等集展开分类研究.首先,对最暗源星等集使用10折嵌套交叉验证,然后使用支持向量机(Support Vector Machine,SVM)、随机森林(Random Forest,RF)、XGBoost(eXtreme Gradient Boosting)等算法建立基分类器模型;使用梯度提升树(Gradient Boosting Decision Tree,GBDT)作为元分类器模型.最后,使用基于星系的分类正确率等指标,与功能树(Function Tree,FT)、SVM、RF、GBDT、XGBoost、堆叠降噪自编码(Stacked Denoising AutoEncoders,SDAE)、深度置信网络(Deep Belief Network,DBN)、深度感知决策树(Deep Perception Decision Tree,DPDT)等模型进行分类结果对比分析.实验结果表明,Stacking集成学习模型在最暗源星等集分类中要比FT算法的星系分类正确率提高了将近10%.同其他传统的机器学习算法、较强的提升算法、深度学习算法相比,Stacking集成学习模型也有较大的提升.