6 resultados para Linear discriminant analysis

em ArchiMeD - Elektronische Publikationen der Universität Mainz - Alemanha


Relevância:

100.00% 100.00%

Publicador:

Resumo:

Die Untersuchungen umfassen die Periode 1981 – 2000 und basieren hauptsächlich auf Daten des Deutschen Wetterdienstes (DWD). Relativwerte der Globalstrahlung beziehen sich auf die Rayleigh-Atmosphäre. Das Regressionsmodell nach Angström ermöglicht die Erweiterung des Meßnetzes. In linearer und nichtlinearer Regression und Korrelation ist die Globalstrahlung entweder abhängige (Sonnenscheindauer, Bewölkung) oder unabhängige Variable (Lufttemperatur, Bodentemperatur). Ihre Intensität in Abhängigkeit von Großwetterlagen, Großwettertypen und Luftmassen wird diskutiert. Diesbezüglich werden mit der Linearen Diskriminanzanalyse ähnliche Großwetterlagen und Stationen in signifikant unterschiedenen Gruppen zusammengefaßt, getrennt nach Sommer- und Winterhalbjahr. Abhängig von der Zeit betrachtet, enthalten Globalstrahlung, direkte und diffuse Sonnenstrahlung, Lufttemperatur, Bewölkung und Niederschlag signifikante zyklische Variationen, die gegebenenfalls klimatologisch relevant sind. Weiteren Aufschluß ergeben deshalb die Zeitreihenanalysen. Autokorrelation-Spektralanalysen (ASA) der genannten Variablen werden in integrierten Spektren dargestellt. Hinweise auf die zeitliche Konstanz signifikanter Varianzmaxima enthalten die Spektren der dynamischen (gleitenden) ASA.

Relevância:

80.00% 80.00%

Publicador:

Resumo:

Das Ziel der Arbeit war die Entwicklung computergestützter Methoden zur Erstellung einer Gefahrenhinweiskarte für die Region Rheinhessen, zur Minimierung der Hangrutschungsgefährdung. Dazu wurde mit Hilfe zweier statistischer Verfahren (Diskriminanzanalyse, Lo­gistische Regression) und einer Methode aus dem Bereich der Künstlichen Intelligenz (Fuzzy Logik) versucht, die potentielle Gefährdung auch solcher Hänge zu klassifizieren, die bis heute noch nicht durch Massenbewegungen aufgefallen sind. Da ingenieurgeologische und geotechnische Hanguntersuchungen aus Zeit und Kostengründen im regionalen Maßstab nicht möglich sind, wurde auf punktuell vorhandene Datenbestände zu einzelnen Rut­schungen des Winters 1981/82, die in einer Rutschungsdatenbank zu­sammengefaßt sind, zurückgegriffen, wobei die daraus gewonnenen Erkenntnisse über Prozeßmechanismen und auslösende Faktoren genutzt und in das jeweilige Modell integriert wurden. Flächenhafte Daten (Lithologie, Hangneigung, Landnutzung, etc.), die für die Berechnung der Hangstabilität notwendig sind, wurden durch Fernerkundungsmethoden, dem Digitalisieren von Karten und der Auswertung von Digitalen Geländemodellen (Reliefanalyse) gewonnen. Für eine weiterführende Untersuchung von einzelnen, als rutschgefährdet klassifi­zierten Bereichen der Gefahrenhinweiskarte, wurde am Beispiel eines Testgebietes, eine auf dem infinite-slope-stability Modell aufbauende Me­thode untersucht, die im Maßstabsbereich von Grundkarten (1:5000) auch geotechnische und hydrogeologische Parameter berücksichtigt und damit eine genauere, der jeweiligen klimatischen Situation angepaßte, Gefahrenabschätzung ermöglicht.

Relevância:

80.00% 80.00%

Publicador:

Resumo:

The relationship and phylogeny of the western Palearctic harvestmen family Trogulidae is investigated. The traditional system of seven genera and approximately 40 species appeared to be artificially composed but a phylogenetic approach and a comprehensive revision has long been sought after. Species are poorly characterised due to their uniform morphology and species evaluation is furthermore complicated by the variability of the few characters used for species delineation. To meet these demands a molecular genetic analysis is accomplished using the nuclear 28S rRNA gene and the mitochondrial cytochrome b gene. This analysis incorporates most genera and species of Trogulidae as well as a comprehensive set of Nemastomatidae and Dicranolasmatidae as outgroup taxa. Phylogenetic results of Bayesian analysis, Maximum Parsimony, Maximum Likelihood and Neighbor Joining are compared with distributional data, morphological characters and results of canonical discriminant analysis of morphometric characters and general congruence of these data sets is shown. To demonstrate the applicability of this method the revision of two species-groups within Trogulus is set out in detail. The Trogulus hirtus species-group and the Trogulus coriziformis species-group are revised. The former is in the central and north-western Balkan Peninsula. T. tricarinatus ssp. hirtus is raised to species level and four new species are described (T. karamanorum [man.n.], T. melitensis [man.n.], T. pharensis [man.n]; T. thaleri [man.n.]). The Trogulus coriziformis species-group is confined to the western Mediterranean area. T. coriziformis, T. aquaticus are re-described, T. cristatus and T. lusitanicus are re-established and four species are described as new (T. balearicus, T. huberi, T. prietoi, T. pyrenaicus). In both species-groups two further cryptic species probably exist but were not described. The species groups are shown to represent different phylogenetic levels and this information is used for the revisional work on the genus Trogulus as well as for the generic system of Trogulidae. Family status of Dicranolasmatidae is rejected and Dicranolasma is shown to be best incorporated within Trogulidae. Calathocratus, Platybessobius and Trogulocratus appear to be polyphyletic and are best to be united within Calathocratus, the oldest name of this set. The cryptic diversity within Trogulidae, especially in Trogulus and the composed genus Calathocratus rates to 150-235% and is thereby remarkably high for a group of the generally well researched European fauna. Genetic features of the group such as heteroplasmy, the possibility of major gene rearrangements and usability of the cytochrome b gene for phylogenetic studies in Opiliones are outlined.

Relevância:

80.00% 80.00%

Publicador:

Resumo:

Lava aus der quartären Vulkanregion der Eifel wurde in römischer Zeit großmaßstäblich abgebaut, um daraus Mühlsteine hoher Qualität herzustellen, die in großen Mengen in weite Teile des Römischen Reiches exportiert wurden. Somit erweisen sich diese Mühlsteine als ideale Indikatoren für Handelsgebiete und –wege dieser Zeit. Für eine erfolgreiche Herkunftsanalyse von Fundstücken ist eine genaue Charakterisierung der infrage kommenden Rohstoffe die unbedingte Voraussetzung. Aus diesem Grund konzentriert sich diese Arbeit auf die geochemische Definition der 16 bekannten römischen „Basalt“-Abbaustellen in den Vulkanfeldern der Ost- und Westeifel, mit dem vorrangigen Ziel, eine möglichste eindeutige Charakterisierung der einzelnen Abbaustellen zu erreichen. Auf dieser Basis wird eine Vorgehensweise zur folgenden Provenienzanalyse der Mühlsteine entwickelt. Um die geochemische Variabilität der Laven zu erfassen wurden die Abbaustellen großflächig beprobt. Die Proben wurden zunächst petrographisch bestimmt und der Phasenbestand mittels Röntgendiffraktometrie (XRD) untersucht. Die Haupt- und Spurenelement-Zusammensetzung wurde anhand von Röntgenfluoreszenzanalysen und Laser-Massenspektrometrie mit induktiv gekoppeltem Plasma (LA-ICP-MS) ermittelt, die Sr-Nd-Pb-Isotopie ausgewählter Proben mittels Thermo-Ionen-Massenspektrometrie (TIMS). Es zeigte sich, dass auf Grundlage der Haupt- und Spurenelementzusammensetzung, gemessen mit RFA, eine Abgrenzung der Abbaustellen zueinander mittels einer Kombination aus geochemischer Diskriminierung mit Cluster- und Diskriminanzanalysen gelingt. Die anschließende Provenienzanalyse der Mühlsteinfundstücke erfolgt analog zur Charakterisierung der Abbaustellen. Im ersten Schritt wird eine geochemische Zuordnung des Fundstücks vorgenommen, um zu überprüfen, ob eine Eifeler Herkunft überhaupt in Frage kommt. Lässt der geochemische Vergleich mit den Referenzdaten aus dem Gelände eine Eifeler Herkunft zu, kommt im zweiten Schritt die Clusteranalyse zum Einsatz. Hier wird auf multivariater Ebene geprüft, ob das Fundstück einer Eifeler Abbaustelle oder einem Cluster von Eifeler Abbaustellen zugeordnet werden kann. Bei einer positiven Zuordnung wird ergänzend als dritter Schritt die Diskriminanzanalyse angewendet, anhand der das Mühlstein-Fundstück einer Abbaustelle mit einer bestimmten Wahrscheinlichkeit zugewiesen wird. Bei 68 analysierten Mühlsteinen aus verschiedenen Regionen Mitteleuropas konnte so für 49 Fundstücke eine Eifeler Provenienz festgestellt werden. Alle römischen Eifeler Mühlsteine entstammen dem Bellerberg bei Mayen, ein mittelalterliches Stück aus dem Niedermendiger Lavastrom. Die Zuordnung der übrigen 19 Mühlsteine zu anderen möglichen Abbaugebieten erfolgt anhand geochemischer Daten aus der Literatur. Aufgrund der unzureichenden Datenlage sind im Falle dieser Proben jedoch im Gegensatz zur Lavastrom-genauen Zuordnung der Eifeler Mühlen lediglich Vermutungen über die Herkunftsregionen möglich. Damit zeigt sich, dass eine ausführliche Referenzdatenbank der in Frage kommenden Rohstoffe die erforderliche Basis für erfolgreiche Provenienzanalysen ist. Die Möglichkeit der genauen Zuordnungen von Mühlsteinen zu den Eifeler Abbaustellen belegt, dass die Verbindung von geochemischer Auswertung mit Cluster- und Diskriminanzanalysen ein hilfreiches Verfahren zur Provenienzanalyse darstellt.

Relevância:

80.00% 80.00%

Publicador:

Resumo:

Numerical simulation of the Oldroyd-B type viscoelastic fluids is a very challenging problem. rnThe well-known High Weissenberg Number Problem" has haunted the mathematicians, computer scientists, and rnengineers for more than 40 years. rnWhen the Weissenberg number, which represents the ratio of elasticity to viscosity, rnexceeds some limits, simulations done by standard methods break down exponentially fast in time. rnHowever, some approaches, such as the logarithm transformation technique can significantly improve rnthe limits of the Weissenberg number until which the simulations stay stable. rnrnWe should point out that the global existence of weak solutions for the Oldroyd-B model is still open. rnLet us note that in the evolution equation of the elastic stress tensor the terms describing diffusive rneffects are typically neglected in the modelling due to their smallness. However, when keeping rnthese diffusive terms in the constitutive law the global existence of weak solutions in two-space dimension rncan been shown. rnrnThis main part of the thesis is devoted to the stability study of the Oldroyd-B viscoelastic model. rnFirstly, we show that the free energy of the diffusive Oldroyd-B model as well as its rnlogarithm transformation are dissipative in time. rnFurther, we have developed free energy dissipative schemes based on the characteristic finite element and finite difference framework. rnIn addition, the global linear stability analysis of the diffusive Oldroyd-B model has also be discussed. rnThe next part of the thesis deals with the error estimates of the combined finite element rnand finite volume discretization of a special Oldroyd-B model which covers the limiting rncase of Weissenberg number going to infinity. Theoretical results are confirmed by a series of numerical rnexperiments, which are presented in the thesis, too.

Relevância:

80.00% 80.00%

Publicador:

Resumo:

The study was arranged to manifest its objectives through preceding it with an intro-duction. Particular attention was paid in the second part to detect the physical settings of the study area, together with an attempt to show the climatic characteristics in Libya. In the third part, observed temporal and spatial climate change in Libya was investigated through the trends of temperature, precipitation, relative humidity and cloud amount over the peri-ods (1946-2000), (1946-1975), and (1976-2000), comparing the results with the global scales. The forth part detected the natural and human causes of climate change concentrat-ing on the greenhouse effect. The potential impacts of climate change on Libya were ex-amined in the fifth chapter. As a case study, desertification of Jifara Plain was studied in the sixth part. In the seventh chapter, projections and mitigations of climate change and desertification were discussed. Ultimately, the main results and recommendations of the study were summarized. In order to carry through the objectives outlined above, the following methods and approaches were used: a simple linear regression analysis was computed to detect the trends of climatic parameters over time; a trend test based on a trend-to-noise-ratio was applied for detecting linear or non-linear trends; the non-parametric Mann-Kendall test for trend was used to reveal the behavior of the trends and their significance; PCA was applied to construct the all-Libya climatic parameters trends; aridity index after Walter-Lieth was shown for computing humid respectively arid months in Libya; correlation coefficient, (after Pearson) for detecting the teleconnection between sun spot numbers, NAOI, SOI, GHGs, and global warming, climate changes in Libya; aridity index, after De Martonne, to elaborate the trends of aridity in Jifara Plain; Geographical Information System and Re-mote Sensing techniques were applied to clarify the illustrations and to monitor desertifi-cation of Jifara Plain using the available satellite images MSS, TM, ETM+ and Shuttle Radar Topography Mission (SRTM). The results are explained by 88 tables, 96 figures and 10 photos. Temporal and spatial temperature changes in Libya indicated remarkably different an-nual and seasonal trends over the long observation period 1946-2000 and the short obser-vation periods 1946-1975 and 1976-2000. Trends of mean annual temperature were posi-tive at all study stations except at one from 1946-2000, negative trends prevailed at most stations from 1946-1975, while strongly positive trends were computed at all study stations from 1976-2000 corresponding with the global warming trend. Positive trends of mean minimum temperatures were observed at all reference stations from 1946-2000 and 1976-2000, while negative trends prevailed at most stations over the period 1946-1975. For mean maximum temperature, positive trends were shown from 1946-2000 and from 1976-2000 at most stations, while most trends were negative from 1946-1975. Minimum tem-peratures increased at nearly more than twice the rate of maximum temperatures at most stations. In respect of seasonal temperature, warming mostly occurred in summer and au-tumn in contrast to the global observations identifying warming mostly in winter and spring in both study periods. Precipitation across Libya is characterized by scanty and sporadically totals, as well as high intensities and very high spatial and temporal variabilities. From 1946-2000, large inter-annual and intra-annual variabilities were observed. Positive trends of annual precipi-tation totals have been observed from 1946-2000, negative trends from 1976-2000 at most stations. Variabilities of seasonal precipitation over Libya are more strikingly experienced from 1976-2000 than from 1951-1975 indicating a growing magnitude of climate change in more recent times. Negative trends of mean annual relative humidity were computed at eight stations, while positive trends prevailed at seven stations from 1946-2000. For the short observation period 1976-2000, positive trends were computed at most stations. Annual cloud amount totals decreased at most study stations in Libya over both long and short periods. Re-markably large spatial variations of climate changes were observed from north to south over Libya. Causes of climate change were discussed showing high correlation between tempera-ture increasing over Libya and CO2 emissions; weakly positive correlation between pre-cipitation and North Atlantic Oscillation index; negative correlation between temperature and sunspot numbers; negative correlation between precipitation over Libya and Southern Oscillation Index. The years 1992 and 1993 were shown as the coldest in the 1990s result-ing from the eruption of Mount Pinatubo, 1991. Libya is affected by climate change in many ways, in particular, crop production and food security, water resources, human health, population settlement and biodiversity. But the effects of climate change depend on its magnitude and the rate with which it occurs. Jifara Plain, located in northwestern Libya, has been seriously exposed to desertifica-tion as a result of climate change, landforms, overgrazing, over-cultivation and population growth. Soils have been degraded, vegetation cover disappeared and the groundwater wells were getting dry in many parts. The effect of desertification on Jifara Plain appears through reducing soil fertility and crop productivity, leading to long-term declines in agri-cultural yields, livestock yields, plant standing biomass, and plant biodiversity. Desertifi-cation has also significant implications on livestock industry and the national economy. Desertification accelerates migration from rural and nomadic areas to urban areas as the land cannot support the original inhabitants. In the absence of major shifts in policy, economic growth, energy prices, and con-sumer trends, climate change in Libya and desertification of Jifara Plain are expected to continue in the future. Libya cooperated with United Nations and other international organizations. It has signed and ratified a number of international and regional agreements which effectively established a policy framework for actions to mitigate climate change and combat deserti-fication. Libya has implemented several laws and legislative acts, with a number of ancil-lary and supplementary rules to regulate. Despite the current efforts and ongoing projects being undertaken in Libya in the field of climate change and desertification, urgent actions and projects are needed to mitigate climate change and combat desertification in the near future.