Search

Solar activity is an important driver of long-term climate trends and must be accounted for in climate models. Unfortunately, direct measurements of this quantity over long periods do not exist. The only observation related to solar activity whose records reach back to the seventeenth century are sunspots. Surprisingly, determining the number of sunspots consistently over time has remained until today a challenging statistical problem. It arises from the need of consolidating data from multiple observing stations around the world in a context of low signal-to-noise ratios, non-stationarity, missing data, non-standard distributions and many kinds of errors. The data from some stations experience therefore severe and various deviations over time. In this paper, we propose the first systematic and thorough statistical approach for monitoring these complex and important series. It consists of three steps essential for successful treatment of the data: smoothing on multiple timescales, monitoring using block bootstrap calibrated CUSUM charts and classifying of out-of-control situations by support vector techniques. This approach allows us to detect a wide range of anomalies (such as sudden jumps or more progressive drifts), unseen in previous analyses. It helps us to identify the causes of major deviations, which are often observer or equipment related. Their detection and identification will contribute to improve future observations. Their elimination or correction in past data will lead to a more precise reconstruction of the world reference index for solar activity: the International Sunspot Number., Comment: 27 pages (without appendices), 6 figures

In many applications, a control procedure is required to detect potential deviations in a panel of serially correlated processes. It is common that the processes are corrupted by noise and that no prior information about the in-control data are available for that purpose. This paper suggests a general nonparametric monitoring scheme for supervising such a panel with time-varying mean and variance. The method is based on a control chart designed by block bootstrap, which does not require parametric assumptions on the distribution of the data. The procedure is tailored to cope with strong noise, potentially missing values and absence of in-control series, which is tackled by an intelligent exploitation of the information in the panel. Our methodology is completed by support vector machine procedures to estimate magnitude and form of the encountered deviations (such as stepwise shifts or functional drifts). This scheme, though generic in nature, is able to treat an important applied data problem: the control of deviations in a subset of sunspot number observations which are part of the International Sunspot Number, a world reference for long-term solar activity., Comment: 58 pages, 18 figures

Observing and counting sunspots constitutes one of the longest-running scientific experiment, with first observations dating back to Galileo and the invention of the telescope around 1610. Today the sunspot number (SN) time series acts as a benchmark of solar activity in a large range of physical models. An appropriate statistical modelling, adapted to the time series' complex nature, is however still lacking. In this work, we provide the first comprehensive uncertainty quantification analysis of sunspot counts. Our interest lies in the following three components: the number of spots ($N_s$), the number of sunspot groups ($N_g$), and the composite $N_c$, defined as $N_c:=N_s+10N_g$. Those are reported by a network of observatories around the world, and are corrupted by errors of various types. We use a multiplicative framework to provide, for each of the three components, an estimation of their error distribution in various regimes (short-term, long-term, minima of solar activity). We also propose a robust estimator for the underlying solar signal and fit a density distribution that takes into account intrinsic characteristics such as over-dispersion, excess of zeros, and multiple modes. The estimation of the solar signal underlying the composite $N_c$ may be seen as a robust version of the International Sunspot Number (ISN), a quantity widely used as a proxy of solar activity. Therefore our results on $N_c$ may serve to characterize the uncertainty on ISN as well. Our results paves the way for a future monitoring of the observatories in quasi-real time, with the aim to alert the observers when they start deviating from the network and prevent large drifts from occurring in the network., Comment: 19 pages, 14 figures, journal paper

Deep learning has achieved astonishing results on many tasks with large amounts of data and generalization within the proximity of training data. For many important real-world applications, these requirements are unfeasible and additional prior knowledge on the task domain is required to overcome the resulting problems. In particular, learning physics models for model-based control requires robust extrapolation from fewer samples - often collected online in real-time - and model errors may lead to drastic damages of the system. Directly incorporating physical insight has enabled us to obtain a novel deep model learning approach that extrapolates well while requiring fewer samples. As a first example, we propose Deep Lagrangian Networks (DeLaN) as a deep network structure upon which Lagrangian Mechanics have been imposed. DeLaN can learn the equations of motion of a mechanical system (i.e., system dynamics) with a deep network efficiently while ensuring physical plausibility. The resulting DeLaN network performs very well at robot tracking control. The proposed method did not only outperform previous model learning approaches at learning speed but exhibits substantially improved and more robust extrapolation to novel trajectories and learns online in real-time, Comment: Published at ICLR 2019

Interactions between convective shells in evolved massive stars have been linked to supernova impostors, to the production of the odd-Z elements Cl, K, and Sc, and they might also help generate the large-scale asphericities that are known to facilitate shock revival in supernova explosion models. We investigate the process of ingestion of C-shell material into a convective O-burning shell, including the hydrodynamic feedback from the nuclear burning of the ingested material. Our 3D hydrodynamic simulations span almost 3 dex in the total luminosity $L_\mathrm{tot}$. All but one of the simulations reach a quasi-stationary state with the entrainment rate and convective velocity proportional to $L_\mathrm{tot}$ and $L_\mathrm{tot}^{1/3}$, respectively. Carbon burning provides $14\,$--$\,33\%$ of the total luminosity, depending on the set of reactions considered. Equivalent simulations done on $768^3$ and $1152^3$ grids are in excellent quantitative agreement. The flow is dominated by a few large-scale convective cells. An instability leading to large-scale oscillations with Mach numbers in excess of $0.2$ develops in an experimental run with the energy yield from C burning increased by a factor of 10. This run represents most closely the conditions expected in a violent O-C shell merger, which is a potential production site for odd-Z elements such as K and Sc and which may seed asymmetries in the supernova progenitor. 1D simulations may underestimate the energy generation from the burning of ingested material by as much as a factor two owing to their missing the effect of clumpiness of entrained material on the nuclear reaction rate., Comment: 24 pages, 27 figures, accepted for publication in MNRAS

Collaborations in astronomy and astrophysics are faced with numerous cyber infrastructure challenges, such as large data sets, the need to combine heterogeneous data sets, and the challenge to effectively collaborate on those large, heterogeneous data sets with significant processing requirements and complex science software tools. The cyberhubs system is an easy-to-deploy package for small to medium-sized collaborations based on the Jupyter and Docker technology, that allows web-browser enabled, remote, interactive analytic access to shared data. It offers an initial step to address these challenges. The features and deployment steps of the system are described, as well as the requirements collection through an account of the different approaches to data structuring, handling and available analytic tools for the NuGrid and PPMstar collaborations. NuGrid is an international collaboration that creates stellar evolution and explosion physics and nucleosynthesis simulation data. The PPMstar collaboration performs large-scale 3D stellar hydrodynamics simulation of interior convection in the late phases of stellar evolution. Examples of science that is presently performed on cyberhubs, in the areas 3D stellar hydrodynamic simulations, stellar evolution and nucleosynthesis and Galactic chemical evolution, are presented., Comment: 19 pages, 9 figures, ApJ Supplement Special Issue on Data, accepted

Rapidly accreting white dwarfs (RAWDs) have been proposed as contributors to the chemical evolution of heavy elements in the Galaxy. Here, we test this scenario for the first time and determine the contribution of RAWDs to the solar composition of first-peak neutron-capture elements. We add the metallicity-dependent contribution of RAWDs to the one-zone galactic chemical evolution code OMEGA according to RAWD rates from binary stellar population models combined with metallicity-dependent $i$-process stellar yields calculated following the models of Denissenkov et al. (2017). With this approach we find that the contribution of RAWDs to the evolution of heavy elements in the Galaxy could be responsible for a significant fraction of the solar composition of Kr, Rb, Sr, Y, Zr, Nb, and Mo ranging from $2$ to $45\%$ depending on the element, the enrichment history of the Galactic gas, and the total mass ejected per RAWD. This contribution could explain the missing solar Lighter Element Primary Process for some elements (e.g., Sr, Y, and Zr). We do not overproduce any isotope relative to the solar composition, but $^{96}$Zr is produced in a similar amount. The $i$ process produces efficiently the Mo stable isotopes $^{95}$Mo and $^{97}$Mo. When nuclear reaction rate uncertainties are combined with our GCE uncertainties, the upper limits for the predicted RAWD contribution increase by a factor of $1.5-2$ for Rb, Sr, Y, and Zr, and by 3.8 and 2.4 for Nb and Mo, respectively. We discuss the implication of the RAWD stellar evolution properties on the single degenerate Type Ia supernova scenario., Comment: 19 pages, 13 figures, 3 tables. Submitted to ApJ (version 2)

The stellar yields for galactic modeling applications (SYGMA) code is an open-source module that models the chemical ejecta and feedback of simple stellar populations (SSPs). It is intended for use in hydrodynamical simulations and semi-analytic models of galactic chemical evolution. The module includes the enrichment from asymptotic giant branch (AGB) stars, massive stars, SNIa and neutron-star mergers. An extensive and extendable stellar yields library includes the NuGrid yields with all elements and many isotopes up to Bi. Stellar feedback from mechanic and frequency-dependent radiative luminosities are computed based on NuGrid stellar models and their synthetic spectra. The module further allows for customizable initial-mass functions and supernova Ia (SNIa) delay-time distributions to calculate time-dependent ejecta based on stellar yield input. A variety of r-process sites can be included. A comparison of SSP ejecta based on NuGrid yields with those from Portinari et al. (1998) and Marigo (2001) reveals up to a factor of 3.5 and 4.8 less C and N enrichment from AGB stars at low metallicity, a result we attribute to NuGrid's modeling of hot-bottom burning. Different core-collapse supernova explosion and fallback prescriptions may lead to substantial variations for the accumulated ejecta of C, O and Si in the first $10^7\, \mathrm{yr}$ at $Z=0.001$. An online interface of the open-source SYGMA module enables interactive simulations, analysis and data extraction of the evolution of all species formed by the evolution of simple stellar populations., Comment: 18 pages, 10 figures, 3 tables, published in ApJS

The definition of similarity is a key prerequisite when analyzing complex data types in data mining, information retrieval, or machine learning. However, the meaningful definition is often hampered by the complexity of data objects and particularly by different notions of subjective similarity latent in targeted user groups. Taking the example of soccer players, we present a visual-interactive system that learns users' mental models of similarity. In a visual-interactive interface, users are able to label pairs of soccer players with respect to their subjective notion of similarity. Our proposed similarity model automatically learns the respective concept of similarity using an active learning strategy. A visual-interactive retrieval technique is provided to validate the model and to execute downstream retrieval tasks for soccer player analysis. The applicability of the approach is demonstrated in different evaluation strategies, including usage scenarions and cross-validation tests.

The first peak s-process elements Rb, Sr, Y and Zr in the post-AGB star Sakurai's object (V4334 Sagittarii) have been proposed to be the result of i-process nucleosynthesis in a post-AGB very-late thermal pulse event. We estimate the nuclear physics uncertainties in the i-process model predictions to determine whether the remaining discrepancies with observations are significant and point to potential issues with the underlying astrophysical model. We find that the dominant source in the nuclear physics uncertainties are predictions of neutron capture rates on unstable neutron rich nuclei, which can have uncertainties of more than a factor 20 in the band of the i-process. We use a Monte Carlo variation of 52 neutron capture rates and a 1D multi-zone post-processing model for the i-process in Sakurai's object to determine the cumulative effect of these uncertainties on the final elemental abundance predictions. We find that the nuclear physics uncertainties are large and comparable to observational errors. Within these uncertainties the model predictions are consistent with observations. A correlation analysis of the results of our MC simulations reveals that the strongest impact on the predicted abundances of Rb, Sr, Y and Zr is made by the uncertainties in the (n,$\gamma$) reaction rates of $^{85}$Br, $^{86}$Br, $^{87}$Kr, $^{88}$Kr, $^{89}$Kr, $^{89}$Rb, $^{89}$Sr, and $^{92}$Sr. (abridged), Comment: 30 pages, 10 figures, 5 tables, accepted for publication by The Journal of Physics G

Based on stellar evolution simulations, we demonstrate that rapidly accreting white dwarfs in close binary systems are an astrophysical site for the intermediate neutron-capture process. During recurrent and very strong He-shell flashes in the stable H-burning accretion regime H-rich material enters the He-shell flash convection zone. $^{12}$C(p,$\gamma)^{13}$N reactions release enough energy to potentially impact convection, and i process is activated through the $^{13}$C($\alpha$,n)$^{16}$O reaction. The H-ingestion flash may not cause a split of the convection zone as it was seen in simulations of He-shell flashes in post-AGB and low-Z AGB stars. We estimate that for the production of first-peak heavy elements this site can be of similar importance for galactic chemical evolution as the s-process production by low-mass AGB stars. The He-shell flashes result in the expansion and, ultimately, ejection of the accreted and then i-process enriched material, via super-Eddington luminosity winds or Roche-lobe overflow. The white dwarf models do not retain any significant amount of the accreted mass, with a He retention efficiency of $\leq 10\%$ depending on mass and convective boundary mixing assumptions. This makes the evolutionary path of such systems to supernova Ia explosion highly unlikely., Comment: 6 pages, 4 figures, 1 table, accepted to publication in ApJ Letters after minor changes

The role of compact binary mergers as the main production site of r-process elements is investigated by combining stellar abundances of Eu observed in the Milky Way, galactic chemical evolution (GCE) simulations, binary population synthesis models, and Advanced LIGO gravitational wave measurements. We compiled and reviewed seven recent GCE studies to extract the frequency of neutron star - neutron star (NS-NS) mergers that is needed in order to reproduce the observed [Eu/Fe] vs [Fe/H] relationship. We used our simple chemical evolution code to explore the impact of different analytical delay-time distribution (DTD) functions for NS-NS mergers. We then combined our metallicity-dependent population synthesis models with our chemical evolution code to bring their predictions, for both NS-NS mergers and black hole - neutron star mergers, into a GCE context. Finally, we convolved our results with the cosmic star formation history to provide a direct comparison with current and upcoming Advanced LIGO measurements. When assuming that NS-NS mergers are the exclusive r-process sites, and that the ejected r-process mass per merger event is 0.01 Msun, the number of NS-NS mergers needed in GCE studies is about 10 times larger than what is predicted by standard population synthesis models. These two distinct fields can only be consistent with each other when assuming optimistic rates, massive NS-NS merger ejecta, and low Fe yields for massive stars. Population synthesis models and GCE simulations are in agreement with the current upper limit (O1) established by Advanced LIGO during their first run of observations. Upcoming measurements will provide important constraints on the actual local NS-NS merger rate, will provide insights on the plausibility of the GCE requirement, and will help to define whether or not compact binary mergers can be the dominant source of r-process elements in the Universe., Comment: Accepted in ApJ, 22 pages, 14 figures, 1 table

Galactic chemical evolution is a topic that involves nuclear physics, stellar evolution, galaxy evolution, observation, and cosmology. Continuous communication and feedback between these fields is a key component in improving our understanding of how galaxies form and how elements are created and recycled in galaxies and intergalactic space. In this proceedings, we present the current state of the JINA-NuGrid chemical evolution pipeline. It is designed to probe the impact of nuclear astrophysics uncertainties on galactic chemical evolution, to improve our knowledges regarding the origin of the elements in a cosmological context, and to create the required interdisciplinary connections., Comment: 3 pages, 2 figures, submitted to JPS Conference Proceedings, Nuclei in the Cosmos XIV

We use the OMEGA galactic chemical evolution code to investigate how the assumptions used for the treatment of galactic inflows and outflows impact numerical predictions. The goal is to determine how our capacity to reproduce the chemical evolution trends of a galaxy is affected by the choice of implementation used to include those physical processes. In pursuit of this goal, we experiment with three different prescriptions for galactic inflows and outflows and use OMEGA within a Markov Chain Monte Carlo code to recover the set of input parameters that best reproduces the chemical evolution of nine elements in the dwarf spheroidal galaxy Sculptor. This provides a consistent framework for comparing the best-fit solutions generated by our different models. Despite their different degrees of intended physical realism, we found that all three prescriptions can reproduce in an almost identical way the stellar abundance trends observed in Sculptor. While the three models have the same capacity to fit the data, the best values recovered for the parameters controlling the number of Type Ia supernovae and the strength of galactic outflows, are substantially different and in fact mutually exclusive from one model to another. For the purpose of understanding how a galaxy evolves, we conclude that only reproducing the evolution of a limited number of elements is insufficient and can lead to misleading conclusions. More elements or additional constraints such as the galaxy's star formation efficiency and the gas fraction are needed in order to break the degeneracy between the different modeling assumptions. Our results show that the successes and failures of chemical evolution models are predominantly driven by the input stellar yields, rather than by the complexity of the galaxy model itself. Simple models such as OMEGA are therefore sufficient to test and validate stellar yields., Comment: 16 pages, 9 figures, 1 table, accepted in ApJ

We used a one-zone chemical evolution model to address the question of how many masses and metallicities are required in grids of massive stellar models in order to ensure reliable galactic chemical evolution predictions. We used a set of yields that includes seven masses between 13 and 30 Msun, 15 metallicities between 0 and 0.03 in mass fraction, and two different remnant mass prescriptions. We ran several simulations where we sampled subsets of stellar models to explore the impact of different grid resolutions. Stellar yields from low- and intermediate-mass stars and from Type Ia supernovae have been included in our simulations, but with a fixed grid resolution. We compared our results with the stellar abundances observed in the Milky Way for O, Na, Mg, Si, Ca, Ti, and Mn. Our results suggest that the range of metallicity considered is more important than the number of metallicities within that range, which only affects our numerical predictions by about 0.1 dex. We found that our predictions at [Fe/H] < -2 are very sensitive to the metallicity range and the mass sampling used for the lowest metallicity included in the set of yields. Variations between results can be as high as 0.8 dex, for any remnant mass prescription. At higher [Fe/H], we found that the required number of masses depends on the element of interest and on the remnant mass prescription. With a monotonic remnant mass prescription where every model explodes as a core-collapse supernova, the mass resolution induces variations of 0.2 dex on average. But with a remnant mass prescription that includes islands of non-explodability, the mass resolution can cause variations of about 0.2 to 0.7 dex depending on the choice of metallicity range. With such a prescription, explosive or non-explosive models can be missed if not enough masses are selected, resulting in over- or under-estimations of the mass ejected by massive stars., Comment: 14 pages, 2 tables, 14 figures, submitted to MNRAS

We use a simple one-zone galactic chemical evolution model to quantify the uncertainties generated by the input parameters in numerical predictions, for a galaxy with properties similar to those of the Milky Way. We compiled several studies from the literature to gather the current constraints for our simulations regarding the typical value and uncertainty of seven basic parameters, which are: the lower and upper mass limits of the stellar initial mass function (IMF), the slope of the high-mass end of the stellar IMF, the slope of the delay-time distribution function of Type Ia supernovae (SNe Ia), the number of SNeIa per Msun formed, the total stellar mass formed, and the final mass of gas. We derived a probability distribution function to express the range of likely values for every parameter, which were then included in a Monte Carlo code to run several hundred simulations with randomly selected input parameters. This approach enables us to analyze the predicted chemical evolution of 16 elements in a statistical way by identifying the most probable solutions along with their 68% and 95% confidence levels. Our results show that the overall uncertainties are shaped by several input parameters that individually contribute at different metallicities, and thus at different galactic ages. The level of uncertainty then depends on the metallicity and is different from one element to another. Among the seven input parameters considered in this work, the slope of the IMF and the number of SNe Ia are currently the two main sources of uncertainty., Comment: 20 pages, 12 figures, 7 tables, accepted in ApJ (v.3 - 11/05/2016)

Classical novae are the result of thermonuclear flashes of hydrogen accreted by CO or ONe white dwarfs, leading eventually to the dynamic ejection of the surface layers. These are observationally known to be enriched in heavy elements, such as C, O and Ne that must originate in layers below the H-flash convection zone. Building on our previous work, we now present stellar evolution simulations of ONe novae and provide a comprehensive comparison of our models with published ones. Some of our models include exponential convective boundary mixing to account for the observed enrichment of the nova ejecta even when accreted material has a solar abundance distribution. Our models produce maximum temperature evolution profiles and nucleosynthesis yields in good agreement with models that generate enriched ejecta by assuming that the accreted material was pre-mixed. We confirm for ONe novae the result we reported previously, i.e.\ we found that $^3$He could be produced {\it in situ} in solar-composition envelopes accreted with slow rates ($\dot{M} < 10^{-10}\,M_\odot/\mbox{yr}$) by cold ($T_{\rm WD} < 10^7$ K) CO WDs, and that convection was triggered by $^3$He burning before the nova outburst in that case. In addition, we now find that the interplay between the $^3$He production and destruction in the solar-composition envelope accreted with an intermediate rate, e.g.\ $\dot{M} = 10^{-10}\,M_\odot/\mbox{yr}$, by the $1.15\,M_\odot$ ONe WD with a relatively high initial central temperature, e.g.\ $T_{\rm WD} = 15\times 10^6$ K, leads to the formation of a thick radiative buffer zone that separates the bottom of the convective envelope from the WD surface. (Abridged), Comment: 19 pages, 23 figures, 2 tables, accepted to publication by MNRAS