Automated classification of stellar spectra — II. Two-dimensional classification with neural networks and principal components analysis

We investigate the application of neural networks to the automation of MK spectral classification. The data set for this project consists of a set of over 5000 optical (3800–5200 Å) spectra obtained from objective prism plates from the Michigan Spectral Survey. These spectra, along with their two-dimensional MK classifications listed in the Michigan Henry Draper Catalogue, were used to develop supervised neural network classifiers. We show that neural networks can give accurate spectral type classifications (σ₆₈= 0.82 subtypes, σ_rms= 1.09 subtypes) across the full range of spectral types present in the data set (B2–M7). We show also that the networks yield correct luminosity classes for over 95 per cent of both dwarfs and giants with a high degree of confidence.   Stellar spectra generally contain a large amount of redundant information. We investigate the application of principal components analysis (PCA) to the optimal compression of spectra. We show that PCA can compress the spectra by a factor of over 30 while retaining essentially all of the useful information in the data set. Furthermore, it is shown that this compression optimally removes noise and can be used to identify unusual spectra.   This paper is a continuation of the work carried out by von Hippel et al. (Paper I).     

A Discussion of IGS Solutions and Their Impact on Geodetic and Geophysical Applications

The International Association of Geodesy officially established the International GPS Service (IGS) on Janaury 1, 1994. Its prime objective is to provide support and a rerefence system for a wide variety of scientific and practical applications involving GPS. To fulfill its role the IGS also generates, in addition to its fundamental products (orbital/staion positions and consistent Earth orientation parameters), additional reference-system products providing the necessary infrastructure, standards, and means of calibrations for timing and various atmospheric applications of GPS. The generation and efficient application of IGS products and their impact on a number of positioning and atmospheric applications, including low earth orbit satellites, is reviewed and discussed. @ 1998 John Wiley & Sons, Inc.     

Dynamical stability and evolution of the discs of Sc galaxies

We examine the local stability of galactic discs against axisymmetric density perturbations with special attention to the different dynamics of the stellar and gaseous components. In particular, the discs of the Milky Way and of NGC 6946 are studied. The Milky Way is shown to be stable, whereas the inner parts of NGC 6946, a typical Sc galaxy from the Kennicutt sample, are dynamically unstable. The ensuing dynamical evolution of the composite disc is studied by numerical simulations. The evolution is so fierce that the stellar disc heats up dynamically on a short time-scale to such a high degree, which seems to contradict the morphological appearance of the galaxy. The star formation rate required to cool the disc dynamically is estimated. Even if the star formation rate in NGC 6946 is at present high enough to meet this requirement, it is argued that the discs of Sc galaxies cannot sustain such a high star formation rate for extended periods.     

Semi-automated extraction of digital objective prism spectra

We describe a method for the extraction of spectra from high-dispersion objective prism plates. Our method is a catalogue-driven plate solution approach, making use of the right ascension and declination coordinates for the target objects. In contrast to existing methods of photographic plate reduction, we digitize the entire plate and extract spectra off-line. This approach has the advantages that it can be applied to CCD objective prism images and spectra can be re-extracted (or additional spectra extracted) without having to re-scan the plate. After a brief initial interactive period, the subsequent reduction procedure is completely automatic, resulting in fully reduced, wavelength-justified spectra. We also discuss a method of removing stellar continua using a combination of non-linear filtering algorithms.   The method described is used to extract over 12 000 spectra from a set of 92 objective prism plates. These spectra are used in an associated project to develop automated spectral classifiers based on neural networks.     

Strain determination from the measurement of pebble shapes

Cross-sections of pebbles or other near-ellipsoidal bodies in rocks can be measured in three planes, conveniently chosen orthogonal to each other, but otherwise oriented arbitrarily. The three sets of measurements, referred to a single Cartesian coordinate system, can in principle be used to calculate the average pebble shape. Some of these measurements, however, are necessarily redundant and, generally, not compatible without adjustments for the distribution of an error. The calculation of the average ellipsoid is developed, as is that of the strain which transforms an original sphere into such an ellipsoid. This calculation involves the distribution by a weighted least-squares method of the error due to redundancy and leads to the assignment of an estimated error to each of the second-rank tensor components of the strain. The assumption is made that, after being referred to the common coordinate system, the measured values for each tensor component have a normal distribution. This assumption may not always be correct but could be tested before this method is used.     

Eodiagenetische Vorgänge bei der Bildung von Salinargesteinen

Zusammenfassung Chemische Präzipitate bleiben als Sedimente nur so lange erhalten, als sie mit der darüberstehenden Lösung, dem Bodenwasser des Salinarbeckens, im chemischen Gleichgewicht stehen. Die in dem Laugenkörper solcher Meeresbecken bestehenden, horizontalen Dichtegrenzen, die auf Konzentrationsunterschieden beruhen, können sich durch klimatische, meeresgeographische u. a. Veränderungen vertikal verschieben. Dadurch ist das Sediment (der Bodenkörper) Umbildungen ausgesetzt. Verdünnung der Lauge bringt entweder totale Wiederauflösung oder selektive Rücklösung der leichtestlöslichen Sediment-Anteile mit sich: durch NaCl-Auflösung bei Haliten ein tonig-anhydritisches Residuum (hardground), Sylvinitisierung (bis Halitisierung) durch Ent-Carnallitisierung (MgCl₂-Wegführung) bei Kalisalzen. Tiefgreifende Umkristallisationen und Entstehung von polygonalen Pfeilern erheblicher Tiefenerstreckung sprechen für Stillstand der Sedimentation und für subaquatische Verkarstung. Auf mechanische Umformungen durch Wellenschlag gehen wohl Augensalze und Knollenanhydrite zurück (Tempestite). An paläogeographisch geeigneten Positionen entstanden durch Gleitbewegungen (slumping) Breccienbänke und Falten erheblichen Ausmaßes in Anhydriten und Kalisalzen. Alle diese Vorgänge werden in das früheste Stadium der Diagenese eingeordnet.Zur Vermeidung der aus zwei verschiedenen Sprachen zusammengesetzten Wortkombination Frühdiagenese wird die BezeichnungEodiagense hiermit in Vorschlag gebracht.

---

Chemical deposits are only preserved as sediments when they are in chemical equilibrium with the lowermost part of the brine basin. The brine bedding, caused by density, or concentration respectively, can be disturbed by changes of climate, by currents a. s. o. When the brine boundaries are shifted in vertical direction the bottom sediment may be transformed. A thinning of the brine will give rise either to total re-dissolution or to any selective removal of the highest soluble substance, f. i. NaCl from rock salt, by leaving anhydritic or marly residuals (hardground), or MgCl₂ from Carnallite, by leaving Sylvite etc. Breaks within the progressive sedimentation produce polygonal structures reaching several meters down into the pre-existing halite sediment. And thatone may be recrystallized in big scale. Physical influence, as movement in the waterbody by wind and waves, gave rise to so-called augensalz, fragmental anhydrites a. s. o. Brecciated anhydrites as well as huge creeping folds in anhydrites and in potash beds have to be explaned by slumping.All these transformations are caused by early diagenesis. For avoiding the term early diagenesis≓ which is composed by 2 words of different languages, the wordEodiagenesis is proposed herewith.

---

Résumé Les sédiments chimiques ne sont conservés qu'aussi longtemps qu'ils sont en équilibre chimique avec la partie basale de la saumure qui se trouve au-dessus d'eux. Les plans de stratification, dûs aux différences de densité liées à la concentration en sels dans la saumure peuvent se déplacer verticalement sous l'action d'influences climatiques, océaniques, tectoniques etc., ce qui expose le sédiment déposé à des transformations. Une dilution de la saumure peut provoquer une redissolution totale ou une dissolution sélective des sels suivant leur solubilité. Par example: dans les halites normales, la solubilisation du NaCl laissera comme résidu un »hardground« composé d'une substance argiloanhydritique; dans les sels potassique, par la dissolution sélective du MgCl₂, la carnallitite sera transformée en sylvinite. — Dans nombre de gisements de halite des structures en pilliers polygonaux de quelques mètres de puissance, où la halite peut Être complètement récrystallisée sont probablement à attribuer à un arrÊt de la sédimentation et à une karstification.Des textures »conglomératiques« observées dans les anhydrites et dans les halites (»Knollenanhydrit«, »Augensalz«), sont à rapporter à l'action des vagues et pourraient donc Être des »tempestites«. Des bancs bréchoides et les plis d'ampleur considérable dans les anhydrites et les sels potassiques ne peuvent Être expliqués que par des glissements subaquatiques.Tous ces événements sont à classer dans un stade très précoce de la diagénèse. Pour éviter le terme de »Frühdiagenese« qui provient de deux langues différentes, il est proposé ici, de désigner ces changements diagénétique précoces sous le nom de »éodiagénèse«.

---

, . , . — « » — . , , , - — hardground —, — — . . . . . , , « » «».

     

Thermal metamorphism of primitive meteorites—IX. On the mechanism of trace element loss from Allende heated up to 1400° C

We report data for Ag, Bi, Cd, Co, Cs, Ga, In, Se, Te, Tl and Zn determined by neutron activation in Allende samples heated for one week at 100° increments in the 1000–1400°C range in a low pressure (initially 10^?5atmH₂) environment using an apparatus of novel design. In the extremes, concentrations of these trace elements—initially present at ppm-ppb levels—in unheated material are lowered even farther by factors of 10^?4–10^?5 over a broad temperature span. Loss of some elements above 1000°C extends trends evident below 1000°C; loss of others is even more extreme. On Arrhenius diagrams some elements exhibit but one apparent activation energy over the entire temperature span of loss while others exhibit 2 or 3, each operative in a particular temperature region. These discontinuities seem related to mineralogic/petrologic alteration and probably reflect differences in diffusion mechanism rather than siting differences. The extension of previous experiments into the temperature regime postulated for chondritic differentiation should lead to a better understanding of the evolution of meteoritic parent bodies.     

Zur geomagnetischen Tiefensondierung

Zusammenfassung Eine Kurvenvergleichsmethode zur Interpretation räumlicher Induktions-pfeilprofile wird unter der Voraussetzung entwickelt, dass die Verteilung der von den geomagnetischen Variationen im Erdinnern induzierten Ströme linear ist. Sie gründet sich auf die lineare Abhängigkeit der Länge des Induktionspfeils von der vertikalen Komponente des Biot-Savartschen Gesetzes. Die Frequenzabhängigkeit des Induktionspfeiles ermöglicht die Angabe eines für die Leitfähigkeitsstruktur charakteristischen Widerstandes und einer Induktivität. Wird zur Bestimmung von Induktionspfeilen das gesamte Horizontalfeld benutzt, so geht die Linearität der Beziehung zwischen Pfeillänge und Ortsfunktion verloren. Für eine Reihe von Strommodellen werden derartige Ortsfunktionenp _z berechnet und diskutiert. Sie dienen als Beispiele von Musterkurven für die entwickelte Interpretationsmethode.

---

Summary One curve-matching method is developed in order to interprete the geomagnetic induction arrow profiles when the distribution of telluric currents in the underground is assumed to be linear. It bases on the linear dependence of the lengths of the induction arrows from the vertical component of the Biot-Savarts Law, which is a function of the radius vector. This is valid when the inducing field has only a horizontal component and the induced horizontal field is zero. In case that the model approximates the current distribution one gets a resistivity and an inductivity from the dependence of frequency of the induction arrows. Some model curves concerning several current distributions are drawn and discussed.

---

---

Mitteilung aus dem Geomagnetischen Institut der DAW, Potsdam, Nr. 212.