Filename
stringlengths 22
64
| Paragraph
stringlengths 8
5.57k
|
|---|---|
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | The IRAC 3.6 and 4.5 µm bands primarily trace the old stellar component in a galaxy while the 5.8 and 8.0 µm bands reveal polycyclic aromatic hydrocarbon (PAH) emis- sion from the heated interstellar dust via UV photons ra- diating from massive young stars. The MIPS 24 µm band is dominated by thermal emission from warm (T > 120 K) dust inhabiting the molecular and neutral medium phases of the ISM, while the 70 µm band traces relatively cold (T < 50 K) emitting dust. The relative contribution of the hot dust emission depends on the strength of the radiation field. It has been shown by many semi-empirical dust emis- sion models that an increased radiation field leads to a more rapid increase of the luminosity in the 24 µm band than other IR bands (Li & Draine 2001; Draine & Li 2007). |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | nus Galaxy, in particular, we focus on its star-forming disk and inner spiral arms. This is the first analysis of Circinus using Spitzer images. We compare the mid-infrared images with the gas density of the extended H i disk to unveil star formation sites in the outer disk of Circinus. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | The observations and data reduction are presented in Sections 2 and 3. Sections 4–6 describe the derivation of basic parameters via photometry. We derive the stellar and gas masses of Circinus in Section 7. Identification of local star formation sites and global star formation rates using a range of wavelength calibrations are given in Section 8. The relation between PAH 8 µm emission and 24 µm hot dust emission is explored in Section 9. Finally, a summary and conclusions are given in Section 10. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | The mid-infrared imaging observations of the Circinus Galaxy were carried out with the IRAC (3.6, 4.5, 5.8, and 8.0 µm) and MIPS (24 and 70 µm) bands of Spitzer under the program led by T. Jarrett (PID 50173). The objective was to survey a region large enough to cover the entire H i disk of Circinus, extending well beyond a smaller Spitzer sur- vey of the disk (G. Rieke, PID 40936), with a sensitivity to capture the low surface brightness emission in the extended star-forming disk. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | For both Spitzer instruments, two separate Astronom- ical Observing Request (AOR) observations (epochs) were combined to improve the sensitivity and to mitigate data artifacts and radiation events. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | The IRAC observations consisted of a series of 12 s ex- posures in high dynamic range mode (HDR) using a 20 × 19 grid with 154′′ steps (roughly 50% overlap). Including both epochs, the total field coverage is 50′ × 50′ with a pixel scale of 0.′′6. Each pixel has an exposure depth of roughly two epochs × 12 s × four coverages, i.e., 96 s per pixel. The achieved 1σ sensitivity in surface brightness is 0.059, 0.045, 0.073 and 0.086 MJy sr−1, respectively, for the 3.6, 4.5, 5.8 and 8.0 µm bands. The short-exposure (0.6 s) HDR observa- tions were used to recover the saturated nucleus of Circinus. The MIPS observations consisted of mapping using the medium scan rate with 21 total scan legs each separated by 148′′. Including both epochs, the total area covered is ∼0.9◦× 0.9◦. Each pixel has a depth of coverage or redun- dancy of roughly 50. The achieved 1σ sensitivity in surface brightness is 0.058 and 1.46 MJy sr−1, respectively, for the 24 and 70 µm bands. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | both epochs of the SSC-produced post-BCD (PBD) mosaics, which have adequate quality for science analysis. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | The achieved angular resolutions are 2′′ for all IRAC bands, 6′′ for the MIPS 24 µm band and 18′′ for the MIPS 70 µm band. We summarize the observing parameters in Table 1 and show the coverage footprints of IRAC and MIPS channels in Figure 1. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | High-resolution H i observations of the Circinus Galaxy were carried out with the ATCA, located in Narrabri, Aus- tralia. Three array configurations (375, 750A and 1.5) were used to obtain single pointings. For details see Jones et al. (1999), who analyze the large-scale H i distribution and kinematics of Circinus at an angular resolution of about 1′ (rms noise ∼1.3 mJy beam−1 per 6.6 km s−1 channel). To study the full H i extent of Circinus, which is much larger than the 33′ ATCA primary beam, H i mosaic ob- servations were obtained in the 375-m array. For details see Curran, Koribalski & Bains (2008) who analysed both the large-scale H i emission of Circinus (at a resolution of 124′′ × 107′′ × 4 km s−1) as well as the inner CO disk. The latter was observed with the SEST 15-m dish at a resolution of at 45′′ for CO(1–0) and at 22′′ for CO(2–1) and 1.8 km s−1 channel width. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | As Circinus lies in the Zone of Avoidance, Galactic H i emission is prominent and detected at radial velocities be- tween about –280 and +50 km s−1. H i emission from Circi- nus is observed between +240 and +640 km s−1. Its H i di- ameter is at least a factor of five larger than the largest estimated optical and infrared diameters. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | For the data analysis in this paper we created (1) Galac- tic H i maps and (2) high-resolution H i maps of the inner disk of the Circinus Galaxy for comparison with the Spitzer MIR data. For both tasks we used the miriad software pack- age (Sault, Teuben & Wright 1995). |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | For the detailed comparison of the Spitzer mid-infrared data with the H i data, we need to substantially improve the angular resolution. This can be achieved by robust by robust or uniform weighting of the uv-data, i.e. giving more weight to the longest baselines in the array (.6 km). We use the calibrated ATCA uv-data from Jones et al. (1999) plus H i data taken in the 6A and 6C array configurations to make H i cubes of the inner Circinus disk at an angular resolution of 15′′. At this resolution, achieved with robust = 0 weighting, good signal-to-noise and high image fidelity was obtained. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | tween the small-scale structure and Galactic H i emission over a specific velocity range (see Section 3.2 for details). |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | The location of the Circinus Galaxy, obscured by the Galac- tic Plane, presents a challenge for source characterization. In order to determine its global properties, including its shape, size and photometric properties, as well as study its extended low surface brightness star-forming disk, the foreground Milky Way (MW) stars and Galactic ISM emis- sion must first be identified and subtracted from the Spitzer images. Figure 2 shows the reduced IRAC and MIPS im- ages with the Circinus Galaxy at the center, obscured by thousands of stars and wide-spread dust emission from the Galactic Plane. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | To identify and subtract MW stars, we used detec- tion and characterization tools that were designed for the multi-band imaging observations obtained with the Spitzer and WISE space-borne infrared telescopes, developed by the WISE Science Data Center (WSDC) of the Infrared Processing and Analysis Center (IPAC). The distinguish- ing characteristic of these systems is that they use all bands simultaneously to detect and characterize sources. The de- tection system is described in Marsh & Jarrett (2012) and the source characterization system in Jarrett et al. (2012) and Cutri et al. (2011). For the Circinus observations, we detect ∼90,000 sources, dominated by the IRAC 3.6 and 4.5 µm sensitivity to Galactic stars. Not all of these de- tections are of Galactic origin, many are indeed associated with the Circinus Galaxy. In order to mitigate removal of real emission associated with Circinus, we visually inspected the images, taking special notice of detected sources that were either in close proximity to Circinus spiral arms, or were spatially resolved, or had very red colors: those with colors or spatial properties that were more consistent with association to Circinus rather with the Galactic foreground population, were not removed from the images. Using the WISE source characterization system, the surviving Galac- tic population was then subtracted from the images using the individual IRAC or MIPS point spread function (PSF) to fit the stellar profiles. The resulting images have their MW stellar component removed. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | Figure 1. Spitzer AOR mapping footprints. The first panel shows the IRAC footprints overlayed on a 2MASS Ks-band image. The 2nd panel shows the MIPS-24 map, overlayed on the IRAS 12 µm image, and the third panel shows the MIPS-70 map overlaying the IRAS 100 µm image. The field of view for each panel is about is 1◦ × 1◦, with North pointing up and East pointing to the left. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | Curran, Koribalski & Bains (2008) (see Section 2.2) as well as Parkes H i maps from the Galactic All-Sky Survey (GASS). The velocity range and resolution of these H i data sets allow a clear separation and analysis of both the Galactic features and the extended structure of the Circinus Galaxy. |
Processed_Gas_and_Star_Formation_in_the_Circinus_Galaxy.txt | The high-resolution Spitzer images (see Figure 2) show structures on all scales. While the IRAC 3.6, 4.5, and 5.8 µm images are dominated by scores of foreground stars, and dif- fuse foreground emission (filaments, bubbles, sheets, etc.) is pervasive in the IRAC 8 µm and MIPS 24 µm images. By combining the Parkes and ATCA H i data towards Circi- nus over the Galactic velocity range, we constructed a high angular resolution image that matches the large-scale MIR features. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | noise levels are uncertain. Efficiency could be further in- creased by using posterior gradient information within pro- posals, which Obsidian does not currently support, but which could be emulated using posterior surrogates. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | sistent with observations constitutes an often poorly con- strained inverse problem. One natural way to approach this is forward-modelling, where the responses of various sen- sors on a proposed geological structure are simulated, and the proposed structure is then updated or sampled itera- tively (for examples see Jessell, 2001; Calcagno et al., 2008; Olierook et al., 2015). |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | The incompleteness and uncertainty of the information contained in any single data set frequently mean that there are many possible worlds consistent with the data being an- alyzed (Tarantola and Valette, 1982; Tarantola, 2005). To the extent that information provided by different datasets is com- plementary, combining all available information into a single joint inversion reduces uncertainty in the final results. Ac- complishing this in a principled and self-consistent manner presents several challenges, including: (i) how to weigh con- straints provided by different datasets relative to each other; (ii) how to rule out worlds inconsistent with geological pro- cesses (expert knowledge); (iii) how to present a transparent accounting of the remaining uncertainty; and (iv) how to do all this in a computationally efficient manner. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | Bayesian statistical techniques provide a powerful frame- work for characterizing and fusing disparate sources of probabilistic information (Tarantola and Valette, 1982; Mosegaard and Tarantola, 1995; Sambridge and Mosegaard, 2002; Sambridge et al., 2012). All input sources of infor- mation — from geophysical sensors, geological field ob- servations, previous inferences, or expert knowledge — are treated as probability distributions; this forces the practitioner to make explicit all assumptions, not only about expected values, but about uncertainties. The out- put of a Bayesian method is also a probability distribu- tion (the posterior), for which the gold-standard repre- sentation is a set of samples from a Monte Carlo algo- rithm, in particular Markov chain Monte Carlo (MCMC; Mosegaard and Tarantola, 1995; Sambridge and Mosegaard, 2002). The posterior distribution is a representation of all possible outcomes and hence provides an internal estimate of uncertainty. The uncertainty associated with the posterior can be visualized in terms of the marginal distributions of parameters of interest, or rendered in 3-D voxelisations of information entropy (Wellmann and Regenauer-Lieb, 2012). The posterior also can be readily updated online as new in- formation becomes available, making Bayesian approaches optimal for decision-making under risk and uncertainty. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | the approximation is hard to capture. The irregular shapes and multimodal structure of the posterior distributions for re- alistic geophysics problems makes them hard to explore; the second moment (variance) of the posterior around each local maximum may in these cases significantly underestimate un- certainties. Moreover, the large number of parameters needed to specify 3-D structures also means these irregular poste- riors are embedded in high-dimensional spaces, increasing the computational cost for both optimization and sampling. Therefore, the sampling methods must usually be tailored to each individual problem and no “one-size-fits-all” solution exists. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | These limitations form the backdrop for current work on applying Bayesian principles to 3-D structural modelling. Giraud et al. (2017, 2018) demonstrate an optimization- based Bayesian inversion framework for 3-D geological models, which finds the maximum of the posterior distribu- tion (maximum a posteriori, or MAP), and expresses uncer- tainty in terms of the posterior covariance around the MAP solution; while they show that fusing data reduces uncer- tainty around this mode, they do not attempt to find or char- acterize other modes, or higher moments of the posterior. Ruggeri et al. (2015) investigate several MCMC schemes for sampling a single-sensor inverse problem (crosshole geo- radar travel time tomography), focusing on sequential, lo- calized perturbations of a proposed 3-D model (“sequential geostatistical resampling”, or SGR); they show that sampling is impractically slow due to high dimensionality and corre- lations between model parameters. Laloy et al. (2016) em- bed the SGR proposal within a parallel-tempered sampling scheme to explore multiple posterior modes of a 2-D inverse problem in groundwater flow, improving computational per- formance but not to a cost-effective threshold. The above methods are non-parametric, in that the model parameters simply form a 3-D field of rock properties to which sensors respond. de la Varga and Wellmann (2016); de la Varga et al. (2018) focus on building parametrized 3-D models in order to reduce the problem dimension and to naturally incorporate structural measurements, but have not yet tested this frame- work on a large-scale 3-D joint inversion with multiple sen- sors. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | McCalman et al. (2014a) present Obsidian, a flexible software platform for MCMC sampling of 3-D multi- modal geophysical models on distributed computing clus- ters. Beardsmore et al. (2016) demonstrate Obsidian on a test problem in geothermal exploration, in the Moomba gas field of the Cooper Basin in South Australia, comparing their re- sults to a deterministic inversion of the same area performed by Meixner and Holgate (2009). These papers outline a full- featured open-source inversion method that can fuse hetero- geneous data into a detailed solution, but make few com- ments about how the efficiency and robustness of the method depends on the particular choices they made. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | McCalman et al. (2014a) inversion code. Our interest is in exploring this problem as a case study to determine which aspects of this problem’s posterior present the most signif- icant obstacles to efficient sampling, which updates to the MCMC scheme improve sampling under these conditions, and how plausible alternative choices of problem setup might influence the efficiency of sampling or the robustness of the inversion. The aspects we consider include: correlations be- tween model parameters; relative weights between datasets with poorly constrained uncertainty; and choices of priors representing different possible exploration scenarios. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | In this section we present a brief overview of the Bayesian forward-modeling paradigm to geophysical inversions. We also provide a discussion of implementing Bayesian infer- ence via sampling using MCMC methods. We then present background on the original Moomba inversion problem, commenting on choices made in the inversion process before we begin to explore different choices in subsequent sections. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | defined by a choice of all of the above elements. A non- Bayesian inversion would proceed by minimizing an objec- tive function, one simple form of which is the mean square misfit between the (statistical) model predictions and the data, corresponding to our negative log likelihood (for obser- vational errors that are independent and Gaussian-distributed with precisely known variance). To penalize solutions that are considered a priori unlikely, the objective function might include additional regularization terms corresponding to the negative log priors in our framework. The full objective function would thus correspond to our negative log poste- rior, and minimization of the objective function would cor- respond to maximization of our posterior probability, under some choice of prior. However, regularization does not nec- essarily proceed from a probabilistic interpretation; interpret- ing model elements in terms of probability may motivate different choices of likelihood or prior than the usual non- probabilistic misfit or regularization terms. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | Indeed, there is considerable flexibility in choosing the above elements even in a fully probabilistic context. For ex- ample, the partitioning of information into “data” and “prior knowledge” is neither unique nor cut-and-dried. However, there are guiding principles: the ideal set of parameters θ is both parsimonious — as few as possible to faithfully repre- sent the world — and interpretable, referring to meaningful aspects of the world that can easily be read off the parame- ter vector. Information resulting from processes that can be easily simulated belong in the likelihood: for example, one might argue that the output of a gravimeter should have a Gaussian distribution, because it responds to the mean rock density within a volume and hence obeys the central limit theorem, or that the output of a Geiger counter should fol- low a Poisson distribution to reflect the physics of radioactive decay. Even processes that are not so easily simulated can at least be approximately described, for example by using a mixture distribution to account for outlier measurements (Mosegaard and Tarantola, 1995) or a prior on the unknown noise level in a process Sambridge et al. (2012). Other infor- mation about allowable or likely worlds belongs in the prior, such as the distribution of initial conditions for simulation, or interpretations of datasets with expensive or intractable for- ward models. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | certainty will include all known modes, higher moments of the distribution, or (more simply) providing enough samples from the distribution to characterize it. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | The posterior distribution p(θ|D) is rarely available in closed form. However, it is often known up to a normalizing constant: p(D|θ)p(θ). Sampling methods such as MCMC can therefore be used to approximate the posterior, without having to explicitly evaluate the normalizing constant (the high-dimensional integral in the denominator of Eq. 1). It is to these methods we turn next. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | or rejected and a copy of the previous state added instead (θ[j+1] = θ[j]). This rule guarantees, under certain regu- larity conditions (Chib and Greenberg, 1995), that the se- quence {θ[j]} converges to the required stationary distribu- tion, P (θ|D), in the limit of increasing n. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | This proposal is straightforward to implement, but its effec- tiveness can depend strongly on Σ, and does not in general scale well to rich, high-dimensional world parametrizations. If Σ has too large a scale, the GRW proposal will step too of- ten into regions of low probability, resulting in many repeated states due to rejections; if the scale is too small, the chain will take only small, incremental steps. In both cases, subsequent states are highly correlated. If the shape of Σ is not tuned to capture correlations between different dimensions of θ, the overall scale must usually be reduced to ensure a reasonable acceptance fraction. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | The SGR method (Ruggeri et al., 2015; Laloy et al., 2016) can be seen as a mixture of multivariate Gaussians, in which Σ has highly correlated sub-blocks of parameters, corre- sponding to variations of the world over different spatial scales. Ruggeri et al. (2015) and Laloy et al. (2016) evalu- ate SGR using single-sensor inversions in crosshole geo- radar travel time tomography, with posteriors correspond- ing to a Gaussian process — an unusually tractable (if high- dimensional) problem that could be solved in closed form as a cross-check. These authors found that in general updating blocks of parameters simultaneously was inefficient, which may not be surprising in a high-dimensional model: for a tightly constrained posterior lying along a low-dimensional subspace of parameter space, almost all directions — hence almost all posterior covariance choices — lead towards re- gions of low probability. Directions picked at random with- out regard for the shape of the posterior will scale badly with increasing dimension. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | Many other types of proposals exist, using information from ensembles of particles (Goodman and Weare, 2010), adaptation of the proposal distribution based on the chain’s history (Haario et al., 2001), derivatives of the posterior (Neal et al., 2011; Girolami and Calderhead, 2011), approx- imations to the posterior (Strathmann et al., 2015), and so forth. The GRW proposal is not only easy to write down and fast to evaluate, but requires no derivative information. We will compare and contrast several derivative-free proposals in our experiments below. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | This allows chains with current states spread throughout pa- rameter space to share global information about the pos- terior in such a way that chain i still samples Pi(θ|D) in the long-term limit. The locations of discovered modes dif- fuse from low-βi chains (which can jump freely between relaxed, broadened versions of these modes) towards the β0 = 1 chain, which can then sample from all modes of the unmodified posterior in the correct proportions. The temper- ature ladder should be defined so that adjacent chains on the ladder are sampling from distributions similar enough for swaps to occur frequently. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | Figure 1 illustrates the sampling of a simple bimodal prob- ability distribution (a mixture of two Gaussians) via PTM- CMC. The solid line depicts the true bimodal distribution, while the broken lines shows the stationary distribution of tempered chains for smaller values of β. The tempered chains are more likely to propose moves across modes than the un- tempered chains, and the existence of a sequence of chains ensures that the difference in probability between successive chains is small enough that swaps can take place easily. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | Even without regard to multiple modes, PTMCMC can also help to reduce correlations between successive inde- pendent posterior samples. Laloy et al. (2016) use SGR as a within-chain proposal in a PTMCMC scheme, demonstrat- ing its effects on correlations between samples but noting that the algorithm remains computationally intensive. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | For Metropolis-Hastings MCMC, the acceptance fraction of proposals is easily measured, and for a chain that is performing well should be ∼ 20–50%. Roberts et al. (1997) showed that the optimal acceptance fraction for random walks in the limit of a large number of dimensions is 0.234, which we will take as our target since the proposals we will consider are modified random walks. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | A trace plot of the history of an element of the parameter vector θ over time summarizes the sampling performance at a glance, revealing where in parameter space an algorithm is spending its time; Fig. 2 shows a series of such figures for some of the different MCMC runs in the present work. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | provides a metric for convergence of different chains to the same result, which decreases to 1 as N → ∞. The chains may be stopped and results read out when the metric dips below a target value for all world parameters θ. The precise number of samples needed may depend on the details of the distribution; the metric provides a stopping condition, but not an estimate of how long it will take to achieve. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | The results from this procedure must still be evaluated according to how well the underlying statistical model de- scribes the geophysical data, and whether the results are ge- ologically plausible — although this is not unique to MCMC solutions. The distribution of residuals of model predictions (forward-modeled data sets) from the observed data can be compared to the assumed likelihood. The standard devia- tion or variance of the residuals (relative to the uncertainty) provide a convenient single-number summary, but the spa- tial distribution of residuals may also be important; outliers and/or structured residuals will indicate places where the model fails to predict the data well, and highlight parts of the model parametrization that need refinement. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | Finally, representative instances of the world itself should be visualized to check for surprising features. Given the com- plexity of real-world data, the adequacy of a given model is in part a matter of scientific judgment, or fitness for a par- ticular applied purpose to which the model will be put. We will use the term model inadequacy to refer to model er- rors arising from approximations or inaccuracies in the world parametrization or the mathematical specification of the for- ward model — although there will always be such approxi- mations in real problems, and the presence of model inade- quacy should not imply that the model is unfit for purpose. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | For our experiments we use a customized fork (v0.1.2) of the open-source Obsidian software package. Obsidian was pre- viously presented in McCalman et al. (2014a) and was used to obtain the modeling results of Beardsmore et al. (2016); v0.1.1 was the most recent open-source version publicly available before our work. We refer the reader to previous publications for a comprehensive description of Obsidian, but below we summarize key elements corresponding to the inversion framework set out above. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | where αij is the offset of the mean depth of the top of layer i at site j, and ρis is the rock property of layer i associated with sensor s. Taken together, the rock properties for each layer and the control points for the boundaries between the layers fully specify the world. This parametrization requires that interface depths be single-valued, not for example per- mitting the surface to fold above or below. Such a limita- tion still enables reasonable representations of sedimentary basins, but may hinder faithful modeling of other kinds of structures. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | MCMC: The sampling algorithm used by Obsidian is an adaptive form of PTMCMC, described in detail in Miasojedow et al. (2013). This algorithm allows for the pro- gressive adjustment of the step size used for proposals within each chain, as well as the temperature ladder used to sample across chains, as sampling progresses. A key feature of the adjustment process is that the maximum allowed change to any chain property diminishes over time, made inversely pro- portional to the number of samples; this is necessary to en- sure that the chains converge to the correct distribution in the limit of large numbers of samples (Roberts and Rosenthal, 2007). The Obsidian implementation of PTMCMC also al- lows it to be run on distributed computing clusters, making it truly parallel in resource use as well as in the requirement for multiple chains. |
Processed_Efficiency_and_robustness_in_Monte_Carlo_sampling_.txt | Development (DSD) in South Australia (Beardsmore et al., 2016). Data used in the inversion include Bouguer anomaly; total magnetic intensity; magnetotelluric sensor data; temper- ature measurements from gas wells; and petrophysical labo- ratory measurements based on 115 core samples from holes drilled throughout the region. Rock properties measured for each sample include density, magnetic susceptibility, thermal conductivity, thermal productivity, and resistivity. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | Publications of the Astronomical Society of Australia (PASA) c(cid:13) Astronomical Society of Australia 2015; published by Cambridge University Press. doi: 10.1017/pas.2015.xxx. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | Abstract A quasar catalogue is presented with a total of 510764 objects including 424748 type 1 QSOs and 26623 type 1 AGN complete from the literature to 25 January 2015. Also included are 25015 high-confidence SDSS-based photometric quasars with radio/X-ray associations, 1595 BL Lac objects, and 32783 type 2 objects. Each object is displayed with arcsecond-accurate astrometry, red and blue photometry, redshift, citations, and radio and X-ray associations where present. Also, 114 new spectroscopically confirmed quasars are presented. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | We have entered an era of large data where accu- racy and completeness are the expected norm in every data-driven endeavour. If the project includes legacy data, however, then particular challenges are faced to bring that legacy data up to today’s standard. So it is with quasar cataloguing which includes 20th century legacy data having paper-specific astrometry and pho- tometry standards of great variance. These were cata- logued as reported, most recently in the V´eron-Cetty & V´eron Catalogue of Quasars and Active Nuclei 13th edition (VCV: V´eron-Cetty & V´eron 2010), although the V´erons worked to improve data quality throughout their releases with good success; in their last five re- leases I contributed a small assist by auditing their pre- release data to find any anomalies. However, arcsecond- accurate astrometry was not achieved for some thou- sands of objects and classifications were unreliable for 3% of objects as will be detailed below. Thus the legacy quasar data needed to be cleaned to today’s standard to integrate that data into the databases of today. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | candidates. Over the past few years I’ve worked to re- solve all issues remaining in the legacy data to where I can now present 100% reliable astrometry and 99% re- liable classifications for all objects. Resolved issues are discussed below and outcomes tabulated. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | The HMQ catalogue presents a complete survey of type 1 QSOs/AGN from the literature; type 2 ob- jects are included as a convenience to the user but only at ≈90% completeness from the literature because they are such a heterogeneous group from many crite- ria used throughout 50 years of discovery; nowadays type 2 is used in a practical way to show narrower emission lines than type 1. This catalogue signals a shift from the VCV philosophy of including question- able objects (so flagged by their discoverers) as quasars, so such objects are dropped. BL Lac objects are in- cluded to continue VCV’s tradition of including them. Also some high-confidence photometric quasar candi- dates with radio/X-ray associations are included as dis- cussed in section 8. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | Accordingly, the catalogue presented here brings the community’s quasar archive to a high standard of us- ability where each quasar is recorded with arcsecond- accurate astrometry (out to 2 arcsec for some legacy objects), its original name and discovery citation, red and blue photometry, best redshift with citation, and radio/X-ray associations. The catalogue is available in both flat file and FITS formats3 and has a ReadMe which describes the contents and source data therein. Table 1 shows a few sample lines of the flat file with some explanation of the columns, but the HMQ ReadMe gives full details of the layout and contents. Figure 1 shows the sky coverage of the HMQ; the high density blocks of coverage are primarily from the SDSS; the high density strip at decl -30◦ is from the 2dF QSO Redshift Survey (2QZ: Croom et al. 2004). |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | The sections below describe the issues involved in constructing this catalogue, including data acquisition from the legacy VCV catalogue and the SDSS cata- logues which supply most quasars, quasars newly pub- lished here, and the inclusion of high-confidence photo- metric quasars. AGN classification is detailed. Photom- etry and radio/X-ray associations are summarized with references given. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | dynamic databases like NED4 and SIMBAD5, and in service of that I presented the true position or status of 449 VCV objects which hitherto had no obvious optical counterpart. I’ve continued that task to where the VCV catalogue is now fully mapped, line for line, into the op- tical J2000. This “VCV Optical catalogue”6 identifies and resolves all anomalies including those reported in Flesch (2013). Following is a summary of those issues, and lists of objects supplementary to those presented in Flesch (2013). |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | TYPE: 1st char is the object classification: Q=QSO, A=AGN, B=BL Lac type, q=photometric, K=type 2; see Section 5. Extra chars summarize the associations displayed: R=radio, X=X-ray, 2=double radio lobes. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | COM: comment on photometry: p=POSS-I magnitudes, so blue is POSS-I O, j=SERC Bj, g=SDSS green, +=optically variable, m=nominal proper motion. • PSF: for red & blue sources: ’-’=stellar, 1=fuzzy, n=no psf available, x=not seen in this colour. • REF & ZREF: citations for name and redshift; citations are indexed in the file ”HMQ-references.txt”. • LOBE or EXTRA: if TYPE shows a ’2’ (=lobes), then double radio lobe identifiers are displayed here. Otherwise, any additional radio and/or X-ray identifiers are displayed here. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | these SA57 quasars which however never ap- peared. Edwards/Beauchemin/Borra (1988) resur- veyed SA57, confirming no Gaston object, and re- lated that Hoag in fact “did not confirm many of the Gaston candidates”. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | A total of 61 duplications are found in VCV, of which 39 were presented in Flesch (2013) and one in Cupani et al. (2011). The remaining 21 newly-found duplications are listed in Table 2 which displays the VCV names of the wrongly-sited duplicate and correctly-sited master, the offset between them in arcseconds, and the final HMQ name which is usually the original form. One of these, #10 on the list, deserves a special mention in that VCV took the redshift of the master from SDSS-DR6, but that redshift was corrected by SDSS-DR7 to that given by the original discoverers. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | As a further supplement to Flesch (2013), Table 3 lists 31 more moves of 30+ arcseconds for VCV ob- jects. 29 of these are newly presented, whilst two, #s 4 & 5, are moved from the position given in Flesch (2013) due to better information. Also, four objects were wrongly moved in Flesch (2013), being Q 0004- 019, IRAS 12202+1646, and MS 22225+2114 which are returned to their VCV location, and IXO 32 which is moved to its finding chart location at 09 10 20.2 +07 05 55 (J2000). The right-hand column is a comment on the move, “OA”=original author. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | Also, Flesch (2013) identified 30 VCV objects to de- list because they either are not quasars or have informa- tion of such poor quality that the object is lost beyond any mechanism of recovery. In a similar vein I present 14 VCV objects in Table 4 which are not seen at all, and the methods of Flesch (2013) find no candidate. These objects are obviously not included in the HMQ catalogue. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | HMQ has sourced much data directly from large source catalogues like SDSS, 2QZ (Croom et al. 2004), etc., so the actual uptake from VCV consists of a residue from smaller papers which amount to 19127 type 1 ob- jects, 1083 BL Lac objects, and 1594 type 2 objects for a total of 21794 legacy objects. Most of the citations for these legacy objects have been copied from the VCV citation list. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | In use of the SDSS, it’s not enough to simply take up all SDSS “quasars” because they aren’t as uniformly reliable as that. The SDSS project uses an automated pipeline which matches spectra to standard templates by which any spectrum well-matched to a quasar tem- plate is securely classified as a quasar. But of course many spectra are ambiguous fits to templates, espe- cially for faint objects with noisy spectra. To cater for this, the SDSS consortium periodically issues manually- checked quasar catalogues which are much more reliable than the automated ones, the most recent being SDSS- DR12Q (Pˆaris et al. 2015). I use those manual cata- logues as a definitive source although they do have a few false objects to remove – indeed, each edition typ- ically drops some objects from previous editions. Ob- jects not accepted into those manual catalogues (i.e., “pipeline-only” objects) may not be quasars or their spectra could be too low quality (i.e.,“low significance”) to be credible, but often they simply fell on the wrong side of threshold cutoffs such as magnitude or lumi- nosity. Inspection of the spectra of such pipeline-only quasars shows that indeed many are plainly recogniz- able as quasar spectra, and many are not. Therefore I needed to find criteria to bulk-separate the high signif- icance pipeline-only quasars from the low, in order to accept the good ones into this comprehensive catalogue. Fortunately this could be done as follows. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | As a background, the SDSS data releases 1 to 7 com- prise the “legacy” period in which the classifications of quasars, galaxies, stars, and “unknowns” were released as separate incremental tables, and the manual cata- logues were quasars only. There was no means of bulk- grading the pipeline-only quasars in the legacy data. The current “SDSS3” data releases 8 to 12 are made as single tables which include all classifications and his- torical data, and the manual catalogues include a “su- perset” of all manually inspected classifications. Key enhancements of the SDSS3 releases are the provision of a ZWARNING field which flags problems with the spectra, and a SUBCLASS field which displays values like “BROADLINE” or “STARFORMING”, but is of- ten blank. For the ZWARNING field, I follow docu- mented recommendations and accept only objects with ZWARNING of 0 (no warnings) or 16 (saturation) as well as 256 (poor astrometry for plughole) and some 128 (unplugged, but flag is documented to be wrong for about 200 objects which have strong signal). For the SUBCLASS field, I find from inspection of hundreds of spectra that the better quality spectra bear a popu- lated SUBCLASS field, and the pipeline classifications and redshifts for those objects are of a good significance. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | tbl∗ # Name 3 ChOPS J04206+3245 042348.6+324522 1 3 2 ESO 374-G25 1 3 Q 1257+2843 1 4 Q 0112-27 1 5 Q 0752+617 3 6 3 7 MCG -03.27.026 3 F 265 8 9 1 EXO 1429.9+3717 3 10 F 917 3 11 G 2344-3852 12 3 1 13 Q 1409+732 3 14 3 15 MCG -03.04.046 3 16 MCG -01.02.013 3 17 MCG -03.02.027 1 18 Q 0139-371 19 MCG -01.25.049 3 3 20 MCG -01.40.001 21 3 3 22 MCG -01.30.005 3 23 MCG -02.27.009 3 24 3 25 MCG -02.08.039 3 26 G J0639-5125 1 27 Q 21570-302 3 28 MCG -02.11.014 3 F 296 29 1 30 2E 1416+2523 31 Q 1606.5+29.1 3 ∗ VCV table 1 listed QSOs, table 3 listed AGN. |
Processed_The_Half_Million_Quasars_(HMQ)_Catalogue.txt | J2000 # 023947.5-013512 1 031243.0-551157 2 040126.3-080200 3 053850.6-440508 4 082817.2-704859 5 090517.9-335016 6 095907.7+022816 7 095926.0+011546 8 114741.2+001455 9 130520.1+290247 10 130609.1+285028 11 141107.0+424936 12 141630.7+520407 13 215242.4+134150 14 ∗ VCV table 1 listed QSOs, table 3 listed AGN. † Vmag from VCV, can be blank. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | Parts of Australia have been privileged to see dazzling lights in the night sky as the Aurora Australis – known as the southern lights – puts on a show this year. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | A recent surge in solar activity caused spectacular auroral displays across the world. While common over the polar regions, aurorae are rare over Australia and are typically restricted to far southern regions, such as Tasmania and Victoria. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | But recently, aurorae have been visible over the whole southern half of Australia, seen as far north as Uluru and Brisbane. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | It’s a phenomenon that has existed since the Earth’s formation and has been witnessed by cultures around the world. These cultures developed their own explanation for the lights in the sky – many of which are strikingly similar. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | From a scientific point of view, aurora form when charged particles of solar wind are channelled to the polar regions by Earth’s magnetic field. These particles ionize oxygen and nitrogen molecules in the upper atmosphere, creating light. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | Auroral displays can show various colours, from white, to yellow, red, green, and blue. They can appear as a nebulous glowing arcs or curtains waving across the sky. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | Aurorae are also reported to make strange sounds on rare occasions. Witnesses describe it as a crackling sound, like rustling grass or radio static. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | In the Arctic, the Inuit say the noise is made by spirits playing a game or trying to communicate with the living. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | In 1851, Aboriginal people near Hobart said an aurora made noise like “people snapping their fingers”. The cause of this noise is unknown. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | Aurorae are significant in Australian Indigenous astronomical traditions. Aboriginal people associate aurorae with fire, death, blood, and omens, sharing many similarities with Native American communities. They are quite different from Inuit traditions of the Aurora Borealis, which are more festive. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | Aboriginal people commonly saw aurorae as fires in the cosmos. To the Gunditjmara of western Victoria, they’re Puae buae (“ashes”). To the Gunai of eastern Victoria, they’re bushfires in the spirit world and an omen of a coming catastrophe. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | The Dieri and Ngarrindjeri of South Australia see aurora as fires created by sky spirits. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | As far north as southwestern Queensland, Aboriginal people saw the phenomenon as “feast fires” of the Oola Pikka —- ghostly beings who spoke to Elders through the aurora. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | The Maori of Aotearoa/New Zealand saw aurorae (Tahunui-a-rangi) as the campfires of ancestors reflected in the sky. These ancestors sailed southward in their canoes and settled on a land of ice in the far south. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | The southern lights let people know they will one day return. This is similar to an Algonquin story from North America. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | Mungan Ngour, a powerful sky ancestor in Gunai traditions, set rules for male initiation and put his son, Tundun, in charge of the ceremonies. When people leaked secret information about these ceremonies, Mungan cast down a great fire to destroy the Earth. The people saw this as an aurora. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | Near Uluru, a group of hunters broke Pitjantjatjara law by killing and cooking a sacred emu. They saw smoke rise to the south, towards the land of Tjura. This was the aurora, viewed as poisonous flames that signalled coming punishment. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | The Dieri also believe an aurora is a warning that someone is being punished for breaking traditional laws, which causes great fear. The breaking of traditional laws would result in an armed party coming to kill the lawbreakers when they least expect it. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | In this context, fear of an aurora was utilised to control behaviour and social standards. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | To Aboriginal communities across New South Wales, Victoria, and South Australia, auroral displays represented blood that was shed by warriors fighting a great battle in the sky, or by spirits of the dead rising to the heavens. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | A total lunar eclipse turns the moon red (sometimes called a blood-moon), which was seen by some communities as the spirit of a dead man rising from his grave. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | In 1859, Aboriginal people in South Australia witnessed an auroral display and a total lunar eclipse. This caused great fear an anxiety, signalling the arrival of dangerous spirit beings. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | There could be a repeat of this astronomical double-act as a lunar eclipse will be visible across Australia on Saturday April 4, 2015. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | Hamacher, D.W. (2013). Aurorae in Australian Aboriginal traditions. Journal of Astronomical History & Heritage, 16(2), 207-219. |
Processed_Fire_in_the_sky:_The_southern_lights_in_Indigenous.txt | Hamacher, D.W. & Norris, R.P. (2011). Eclipses in Australian Aboriginal Astronomy. Journal of Astronomical History & Heritage, 14(2), 103-114. |
Processed_Study_of_Quark_Confinement_in_Baryons_with_Lattice.txt | In SU(3) lattice QCD, we perform the detailed study for the ground-state three-quark (3Q) potential V g.s. 3Q and the 1st excited-state 3Q potential V e.s. 3Q , i.e., the energies of the ground state and the 1st excited state of the gluon field in the presence of the static three quarks. From the accurate calculation for more than 300 different patterns of 3Q systems, the static ground-state 3Q potential V g.s. 3Q is found to be well described by the Coulomb plus Y-type linear potential (Y-Ansatz) within 1%-level deviation. As a clear evidence for Y-Ansatz, Y-type flux- tube formation is actually observed on the lattice in maximally-Abelian projected QCD. For about 100 patterns of 3Q systems, we calculate the 1st excited-state 3Q potential V e.s. 3Q , and find a large gluonic-excitation energy ∆E3Q ≡ V e.s. 3Q of about 1 GeV, which gives a physical reason of the success of the quark model even without gluonic excitations. We present also the first study for the penta-quark potential V5Q in lattice QCD, and find that V5Q is well described by the sum of the OGE Coulomb plus multi-Y type linear potential. |
Processed_Study_of_Quark_Confinement_in_Baryons_with_Lattice.txt | Quantum chromodynamics (QCD), the SU(3) gauge theory, was first proposed by Yoichiro Nambu [1] in 1966 as a candidate for the funda- mental theory of the strong interaction, just af- ter the introduction of the “new” quantum num- ber, “color” [2]. In spite of its simple form, QCD creates thousands of hadrons and leads to vari- ous interesting nonperturbative phenomena such as color confinement [3] and dynamical chiral- symmetry breaking [4]. Even now, it is very diffi- cult to deal with QCD due to its strong-coupling nature in the infrared region. |
Processed_Study_of_Quark_Confinement_in_Baryons_with_Lattice.txt | role of the Q- ¯Q potential for meson properties, and both the Q- ¯Q potential and the 3Q poten- tial are equally important fundamental quanti- ties in QCD. Furthermore, the 3Q potential is the key quantity to clarify the quark confinement in baryons. However, in contrast to the Q- ¯Q po- tential [5], there was almost no lattice QCD study for the 3Q potential before our study in 1999 [9], in spite of its importance in the hadron physics. |
Processed_Study_of_Quark_Confinement_in_Baryons_with_Lattice.txt | In recent years, the lattice QCD Monte Carlo calculation becomes a reliable and useful method for the analysis of nonperturbative QCD [5], which indicates an important direction in the In this paper, using lattice hadron physics. QCD, we study the inter-quark potential in de- tail [6,7,8,9]. |
Processed_Study_of_Quark_Confinement_in_Baryons_with_Lattice.txt | As for the 3Q potential form, we note two theoret- ical arguments at short and long distance limits. |
Processed_Study_of_Quark_Confinement_in_Baryons_with_Lattice.txt | where Lmin is the minimal value of the total flux- tube length. Of course, it is nontrivial that these simple arguments on UV and IR limits of QCD hold for the intermediate region. Then, we study the 3Q potential in lattice QCD. Note that the lattice QCD calculation is completely indepen- dent of any Ansatz for the potential form. |
Processed_Study_of_Quark_Confinement_in_Baryons_with_Lattice.txt | dxµAµ(x)} in Fig.1. The with Uk ≡ P exp{ig RΓk 3Q Wilson loop physically means that a color- singlet gauge-invariant 3Q state is created at t = 0 and is annihilated at t = T with the three quarks spatially fixed for 0 < t < T . |
Processed_Study_of_Quark_Confinement_in_Baryons_with_Lattice.txt | where Vn denotes the nth energy of the gauge- field configuration in the presence of the spatially- fixed three quarks [6,7,8]. |
Processed_Study_of_Quark_Confinement_in_Baryons_with_Lattice.txt | While Vn depends only on the 3Q location, Cn depends on the operator choice for the 3Q state. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.