What is flagellar swimming? Cilia and flagella are whip-like cell appendages that can exhibit regular bending waves. This active process emerges from the non-equilibrium dynamics of molecular motors distributed along the length of cilia and flagella. Eukaryotic cells can possess many cilia and flagella that beat in a coordinated fashion, thus transporting fluids, as in mammalian airways or the ventricular system inside the brain. Many unicellular organisms posses just one or two flagella, rendering them microswimmers that are propelled through fluids by the flagellar beat including sperm cells and the biflagellate green alga Chlamydomonas. Objectives. In this thesis in theoretical biological physics, we seek to understand the nonlinear dynamics of flagellar swimming and synchronization. We investigate the flow fields induced by beating flagella and how in turn external hydrodynamic flows change speed and shape of the flagellar beat. This flagellar load-response is a prerequisite for flagellar synchronization. We want to find the physical principals underlying stable synchronization of the two flagella of Chlamydomonas cells. Results. First, we employed realistic hydrodynamic simulations of flagellar swimming based on experimentally measured beat patterns. For this, we developed analysis tools to extract flagellar shapes from high-speed videoscopy data. Flow-signatures of flagellated swimmers are analysed and their effect on a neighboring swimmer is compared to the effect of active noise of the flagellar beat. We were able to estimate a chemomechanical energy efficiency of the flagellar beat and determine its waveform compliance by comparing findings from experiments, in which a clamped Chlamydomonas is exposed to external flow, to predictions from an effective theory that we designed. These mechanical properties have interesting consequences for the synchronization dynamics of Chlamydomonas, which are revealed by computer simulations. We propose that direct elastic coupling between the two flagella of Chlamydomonas, as suggested by recent experiments, in combination with waveform compliance is crucial to facilitate in-phase synchronization of the two flagella of Chlamydomonas.:1 Introduction 1.1 Physics of cell motility: flagellated swimmers as model system 2 1.1.1 Tissue cells and unicellular eukaryotic organisms have cilia and flagella 2 1.1.2 The conserved architecture of flagella 3 1.1.3 Synchronization in collections of flagella 5 1.2 Hydrodynamics at the microscale 9 1.2.1 Navier-Stokes equation 10 1.2.2 The limit of low Reynolds number 10 1.2.3 Multipole expansion of flow fields 11 1.3 Self-propulsion by viscous forces 13 1.3.1 Self propulsion requires broken symmetries 13 1.3.2 Signatures of flowfields: pusher & puller 15 1.4 Overview of the thesis 16 2 Flow signatures of flagellar swimming 2.1 Self-propulsion of flagellated swimmers 20 2.1.1 Representation of flagellar shapes 20 2.1.2 Computation of hydrodynamic friction forces 21 2.1.3 Material frame and rigid-body transformations 22 2.1.4 The grand friction matrix 23 2.1.5 Dynamics of swimming 23 2.2 The hydrodynamic far field: pusher and puller 26 2.2.1 The flow generated by a swimmer 26 2.2.2 Force dipole characterization 27 2.2.3 Flagellated swimmers alternate between pusher and puller 29 2.2.4 Implications for two interacting Chlamydomonas cells 31 2.3 Inertial screening of oscillatory flows 32 2.3.1 Convection and oscillatory acceleration 33 2.3.2 The oscilet: fundamental solution of unsteady flow 35 2.3.3 Screening length of oscillatory flows 35 2.4 Energetics of flagellar self-propulsion 36 2.4.1 Impact of inertial screening on hydrodynamic dissipation 37 2.4.2 Case study: the green alga Chlamydomonas 38 2.4.3 Discussion: evolutionary optimization and the number of molecular motors 38 2.5 Summary 39 3 The load-response of the flagellar beat 3.1 Experimental collaboration: flagellated swimmers exposed to flows 41 3.1.1 Description of the experimental setup 42 3.1.2 Computed flow profile in the micro-fluidic device 43 3.1.3 Image processing and flagellar tracking 43 3.1.4 Mode decomposition and limit-cycle reconstruction 47 3.1.5 Changes of limit-cycle dynamics: deformation, translation, acceleration 49 3.2 An effective theory of flagellar oscillations 50 3.2.1 A balance of generalized forces 50 3.2.2 Hydrodynamic friction in generalized coordinates 51 3.2.3 Intra-flagellar friction 52 3.2.4 Calibration of active flagellar driving forces 52 3.2.5 Stability of the limit cycle of the flagellar beat 53 3.2.6 Equations of motion 55 3.3 Comparison of theory and experiment 56 3.3.1 Flagellar mean curvature 57 3.3.2 Susceptibilities of phase speed and amplitude 57 3.3.3 Higher modes and stalling of the flagellar beat at high external load 59 3.3.4 Non-isochrony of flagellar oscillations 63 3.4 Summary 63 4 Flagellar load-response facilitates synchronization 4.1 Synchronization to external driving 65 4.2 Inter-flagellar synchronization in the green alga Chlamydomonas 67 4.2.1 Equations of motion for inter-flagellar synchronization 68 4.2.2 Synchronization strength for free-swimming and clamped cells 70 4.2.3 The synchronization strength depends on energy efficiency and waveform compliance 73 4.2.4 The case of an elastically clamped cell 74 4.2.5 Basal body coupling facilitates in-phase synchronization 75 4.2.6 Predictions for experiments 78 4.3 Summary 80 5 Active flagellar fluctuations 5.1 Effective description of flagellar oscillations 84 5.2 Measuring flagellar noise 84 5.2.1 Active phase fluctuations are much larger than thermal noise 84 5.2.2 Amplitude fluctuations are correlated 85 5.3 Active flagellar fluctuations result in noisy swimming paths 86 5.3.1 Effective diffusion of swimming circles of sperm cell 86 5.3.2 Comparison of the effect of noise and hydrodynamic interactions 87 5.4 Summary 88 6 Summary and outlook 6.1 Summary of our results 89 6.2 Outlook on future work 90 A Solving the Stokes equation A.1 Multipole expansion 95 A.2 Resistive-force theory 96 A.3 Fast multipole boundary element method 97 B Linearized Navier-Stokes equation B.1 Linearized Navier-Stokes equation 101 B.2 The case of an oscillating sphere 102 B.3 The small radius limit 103 B.4 Greens function 104 C Hydrodynamic friction C.1 A passive particle 107 C.2 Multiple Particles 107 C.3 Generalized coordinates 108 D Data analysis methods D.1 Nematic filter 111 D.1.1 Nemat 111 D.1.2 Nematic correlation 111 D.2 Principal-component analysis 112 D.3 The quality of the limit-cycle projections of experimental data 113 E Adler equation F Sensitivity analysis for computational results F.1 The distance function of basal coupling 117 F.2 Computed synchronization strength for alternative waveform 118 F.3 Insensitivity of computed load-response to amplitude correlation time 118 List of Symbols List of Figures Bibliography

Tunneling is one of the most prominent features of quantum mechanics. While the tunneling process in one-dimensional integrable systems is well understood, its quantitative prediction for systems with mixed phase space is a long-standing open challenge. In such systems regions of regular and chaotic dynamics coexist in phase space, which are classically separated but quantum mechanically coupled by the process of dynamical tunneling. We derive a prediction of dynamical tunneling rates which describe the decay of states localized inside the regular region towards the so-called chaotic sea. This approach uses a fictitious integrable system which mimics the dynamics inside the regular domain and extends it into the chaotic region. Excellent agreement with numerical data is found for kicked systems, billiards, and optical microcavities, if nonlinear resonances are negligible. Semiclassically, however, such nonlinear resonance chains dominate the tunneling process. Hence, we combine our approach with an improved resonance-assisted tunneling theory and derive a unified prediction which is valid from the quantum to the semiclassical regime. We obtain results which show a drastically improved accuracy of several orders of magnitude compared to previous studies. / Der Tunnelprozess ist einer der bedeutensten Effekte in der Quantenmechanik. Während das Tunneln in eindimensionalen integrablen Systemen gut verstanden ist, gestaltet sich dessen Beschreibung für Systeme mit gemischtem Phasenraum weitaus schwieriger. Solche Systeme besitzen Gebiete regulärer und chaotischer Bewegung, die klassisch getrennt sind, aber quantenmechanisch durch den Prozess des dynamischen Tunnelns gekoppelt werden. In dieser Arbeit wird eine theoretische Vorhersage für dynamische Tunnelraten abgeleitet, die den Zerfall von Zuständen, die im regulären Gebiet lokalisiert sind, in die sogenannte chaotische See beschreibt. Dazu wird ein fiktives integrables System konstruiert, das im regulären Bereich eine nahezu gleiche Dynamik aufweist und diese Dynamik in das chaotische Gebiet fortsetzt. Die Theorie zeigt eine ausgezeichnete Übereinstimmung mit numerischen Daten für gekickte Systeme, Billards und optische Mikrokavitäten, falls nichtlineare Resonanzketten vernachlässigbar sind. Semiklassisch jedoch bestimmen diese nichtlinearen Resonanzketten den Tunnelprozess. Daher kombinieren wir unseren Zugang mit einer verbesserten Theorie des Resonanz-unterstützten Tunnelns und erhalten eine Vorhersage,die vom Quanten- bis in den semiklassischen Bereich gültig ist. Ihre Resultate zeigen eine Genauigkeit, die verglichen mit früheren Theorien um mehrere Größenordnungen verbessert wurde.

Proceedings of the 18th IEEE Workshop on Nonlinear Dynamics of Electronic Systems, which took place in Dresden, Germany, 26 – 28 May 2010.:Welcome Address ........................ Page I Table of Contents ........................ Page III Symposium Committees .............. Page IV Special Thanks ............................. Page V Conference program (incl. page numbers of papers) ................... Page VI Conference papers Invited talks ................................ Page 1 Regular Papers ........................... Page 14 Wednesday, May 26th, 2010 ......... Page 15 Thursday, May 27th, 2010 .......... Page 110 Friday, May 28th, 2010 ............... Page 210 Author index ............................... Page XIII

We investigate systems with a mixed phase space, where regular and chaotic dynamics coexist. Classically, regions with regular motion, the regular islands, are dynamically not connected to regions with chaotic motion, the chaotic sea. Typically, this is also reflected in the quantum properties, where eigenstates either concentrate on the regular or the chaotic regions. However, it was shown that quantum mechanically, due to the tunneling process, a coupling is induced and flooding of regular islands may occur. This happens when the Heisenberg time, the time needed to resolve the discrete spectrum, is larger than the tunneling time from the regular region to the chaotic sea. In this case the regular eigenstates disappear. We study this effect by the time evolution of wave packets initially started in the chaotic sea and find increasing probability in the regular island. Using random matrix models a quantitative prediction is derived. We find excellent agreement with numerical data obtained for quantum maps and billiards systems. For open systems we investigate the phenomenon of flooding and disappearance of regular states, where the escape time occurs as an additional time scale. We discuss the reappearance of regular states in the case of strongly opened systems. This is demonstrated numerically for quantum maps and experimentally for a mushroom shaped microwave resonator. The reappearance of regular states is explained qualitatively by a matrix model. / Untersucht werden Systeme mit gemischtem Phasenraum, in denen sowohl reguläre als auch chaotische Dynamik auftritt. In der klassischen Mechanik sind Gebiete regulärer Bewegung, die sogenannten regulären Inseln, dynamisch nicht mit den Gebieten chaotischer Bewegung, der chaotischen See, verbunden. Dieses Verhalten spiegelt sich typischerweise auch in den quantenmechanischen Eigenschaften wider, so dass Eigenfunktionen entweder auf chaotischen oder regulären Gebieten konzentriert sind. Es wurde jedoch gezeigt, dass aufgrund des Tunneleffektes eine Kopplung auftritt und reguläre Inseln geflutet werden können. Dies geschieht wenn die Heisenbergzeit, das heißt die Zeit die das System benötigt, um das diskrete Spektrum aufzulösen, größer als die Tunnelzeit vom Regulären ins Chaotische ist, wobei reguläre Eigenzustände verschwinden. Dieser Effekt wird über eine Zeitentwicklung von Wellenpaketen, die in der chaotischen See gestartet werden, untersucht. Es kommt zu einer ansteigenden Wahrscheinlichkeit in der regulären Insel. Mithilfe von Zufallsmatrixmodellen wird eine quantitative Vorhersage abgeleitet, welche die numerischen Daten von Quantenabbildungen und Billardsystemen hervorragend beschreibt. Der Effekt des Flutens und das Verschwinden regulärer Zustände wird ebenfalls mit offenen Systemen untersucht. Hier tritt die Fluchtzeit als zusätzliche Zeitskala auf. Das Wiederkehren regulärer Zustände im Falle stark geöffneter Systeme wird qualitativ mithilfe eines Matrixmodells erklärt und numerisch für Quantenabbildungen sowie experimentell für einen pilzförmigen Mikrowellenresonator belegt.

Generic Hamiltonian systems have a mixed phase space, in which regular and chaotic motion coexist. In the chaotic sea the classical transport is limited by partial barriers, which allow for a flux \Phi given by the corresponding turnstile area. Quantum mechanically the transport is suppressed if Planck's constant is large compared to the classical flux, h >> \Phi, while for h << \Phi classical transport is recovered. For the transition between these limiting cases there are many open questions, in particular concerning the correct scaling parameter and the width of the transition. To investigate this transition in a controlled way, we design a kicked system with a particularly simple phase-space structure, consisting of two chaotic regions separated by one dominant partial barrier. We find a universal scaling with the single parameter \Phi/h and a transition width of almost two orders of magnitude in \Phi/h. In order to describe this transition, we consider several matrix models. While the numerical data is not well described by the random matrix model proposed by Bohigas, Tomsovic, and Ullmo, a deterministic 2x2-model, a channel coupling model, and a unitary model are presented, which describe the transitional behavior of the designed kicked system. This is also confirmed for the generic standard map, suggesting a universal scaling behavior for the quantum transition of a partial barrier. / Generische Hamilton'sche Systeme besitzen einen gemischten Phasenraum, in dem sowohl reguläre als auch chaotische Dynamik vorkommen. Der klassische Transport in der chaotischen See wird durch partielle Barrieren begrenzt, die nur einen Fluss \Phi hindurch lassen. Der quantenmechanische Transport ist stark unterdrückt, wenn die Planck'sche Konstante groß gegen den klassischen Fluss ist, h >> \Phi. Ist hingegen h << \Phi folgt die Quantenmechanik der klassischen Dynamik. Für den Übergangsbereich zwischen diesen Grenzfällen gibt es noch viele offene Fragen, insbesondere bezüglich des richtigen Skalierungsparameters und der Breite des Übergangs. Um gezielt diesen Übergang zu untersuchen, haben wir ein System mit einem besonders einfachen Phasenraum entworfen. Er besteht aus zwei chaotischen Gebieten, die durch eine dominante partielle Barriere getrennt sind. Es zeigt sich, dass das universelle Verhalten durch den Parameter \Phi/h beschrieben wird und der Übergang sich über zwei Größenordnungen erstreckt. Wir betrachten verschiedene Matrixmodelle um diesen Übergang zu verstehen. Die numerischen Daten werden nicht durch das Zufallsmatrixmodell von Bohigas, Tomsovic und Ullmo beschrieben. Ein deterministisches 2x2-Modell, eine Kanalkopplung und ein unitäres Matrixmodell beschreiben hingegen den Übergang des entworfenen gekickten Systems. Die Tatsache, dass auch die generische Standardabbildung diesem Verhalten folgt, spricht für ein universelles Verhalten des Quantenübergangs einer partiellen Barriere.

No description available.

[pt] É um resultado conhecido na literatura que uma cadeia unidimensional de partículas, que interagem harmonicamente com seus primeiros vizinhos, não conduz calor, e forças não lineares são necessárias para reproduzir a lei de Fourier da condução de calor. Quando são introduzidas assimetrias em tal sistema condutor, se obtém um efeito retificador onde a corrente térmica apresenta magnitudes diferentes dependendo de qual lado da cadeia tem maior temperatura, tais dispositivos sendo chamados de diodos térmicos. Neste trabalho estudamos os dois fenômenos, condução de calor e retificação térmica, em uma cadeia unidimensional de partículas, com condições de contorno fixas, acopladas a dois banhos térmicos, um em cada extremidade, modelados como termostatos de Langevin. As partículas interagem com seus primeiros vizinhos harmonicamente e estão sujeitas a um potencial localizado externo não linear, para o qual estudamos dois tipos, os potenciais Frenkel-Kontorova e Ø elevado a 4. Verificamos que a lei de Fourier é observada, para ambos os casos, com o perfil de temperatura e a condutividade térmica dependendo da relação entre as amplitudes harmônica e anarmônica, e a temperatura média do sistema. Em seguida, para criar uma assimetria na cadeia, nós acoplamos dois segmentos de mesmo tamanho. Observamos um efeito retificador onde a direção preferencial difere para cada potencial localizado estudado. A forma como as temperaturas dos banhos térmicos mudam a magnitude da retificação também foi observada. Nós também investigamos o efeito de não linearidades interfaciais, por meio de uma lei de potência que acopla segmentos Ø elevado a 4. Alterando o expoente da lei de potência, nós buscamos as condições sob as quais a retificação ótima é atingida. / [en] It is a known result in the literature that a one-dimensional chain of particles that interact harmonically with its first neighbors does not conduct heat, and nonlinear forces are needed to reproduce Fourier s law of heat conduction. When asymmetries are introduced in such a conducting system, a rectifying effect is obtained where the thermal current shows different magnitudes depending on which side of the chain has higher temperature, such devices being called thermal diodes. In this work we study both phenomena, heat conduction and thermal rectification, in a onedimensional chain of particles, with fixed boundary conditions, coupled to two thermal baths, one at each end, modeled as Langevin thermostats. The particles interact with their first neighbors harmonically and have a nonlinear on-site potential, for which we study two types, Frenkel-Kontorova and Ø 4 potentials. We verify that, for both cases, Fourier s law is observed, where the temperature profile and the thermal conductivity are dependent on the relation between the harmonic and anharmonic amplitudes, and the system s average temperature. Next, to create an asymmetry in the chain, we coupled two different segments of equal lengths. We observed a rectifying effect, where the preferential direction differs for each of the two on-site potentials studied. How the heat-bath temperatures changes the magnitude of rectification was also observed. We also investigated the effect of interfacial nonlinearities through a power-law potential, coupling Ø 4 segments. By changing the power-law exponent, we looked for the conditions under which optimal rectification is achieved.

A flapping-wing micro-air-vehicle (FWMAV) represents a complex multi-disciplinary system whose analysis invokes the frontiers of the aerospace engineering disciplines. From the aerodynamic point of view, a nonlinear, unsteady flow is created by the flapping motion. In addition, non-conventional contributors, such as the leading edge vortex, to the aerodynamic loads become dominant in flight. On the other hand, the flight dynamics of a FWMAV constitutes a nonlinear, non-autonomous dynamical system. Furthermore, the stringent weight and size constraints that are always imposed on FWMAVs invoke design with minimal actuation. In addition to the numerous motivating applications, all these features of FWMAVs make it an interesting research point for engineers. In this Dissertation, some challenging points related to FWMAVs are considered. First, an analytical unsteady aerodynamic model that accounts for the leading edge vortex contribution by a feasible computational burden is developed to enable sensitivity and optimization analyses, flight dynamics analysis, and control synthesis. Second, wing kinematics optimization is considered for both aerodynamic performance and maneuverability. For each case, an infinite-dimensional optimization problem is formulated using the calculus of variations to relax any unnecessary constraints induced by approximating the problem as a finite-dimensional one. As such, theoretical upper bounds for the aerodynamic performance and maneuverability are obtained. Third, a design methodology for the actuation mechanism is developed. The proposed actuation mechanism is able to provide the required kinematics for both of hovering and forward flight using only one actuator. This is achieved by exploiting the nonlinearities of the wing dynamics to induce the saturation phenomenon to transfer energy from one mode to another. Fourth, the nonlinear, time-periodic flight dynamics of FWMAVs is analyzed using direct and higher-order averaging. The region of applicability of direct averaging is determined and the effects of the aerodynamic-induced parametric excitation are assessed. Finally, tools combining geometric control theory and averaging are used to derive analytic expressions for the textit{Symmetric Products}, which are vector fields that directly affect the acceleration of the averaged dynamics. A design optimization problem is then formulated to bring the maneuverability index/criterion early in the design process to maximize the FWMAV maneuverability near hover. / Ph. D.

[pt] O objetivo desta tese é fornecer ferramentas para a análise global de sistemas dinâmicos não determinísticos com atratores coexiostentes considerando incerteza paramétrica ou ruído e aplicá-las a problemas de engenharia. Para isso, é proposta uma estratégia de discretização adaptativa no espaço de fase baseada no método clássico de Ulam. Inicialmente, apresenta-se uma revisão das definições matemáticas de sistemas dinâmicos, incerteza paramétrica e ruído, destacando-se o efeito da aleatoriedade nas estruturas dinâmicas globais. Operadores de transferência discretos são derivados com as modificações necessárias devido à incerteza dos parâmetros. Bacias de atração estocásticas e distribuição dos atratores substituem o conceito usual de bacia e atrator. Para casos de incerteza paramétrica, o espaço de fase é aumentado com o espaço de probabilidade correspondente, resultando em uma coleção de operadores de transferência dos quais médias são obtidas. São propostas duas estratégias de discretização adaptativa no espaço de fase, uma que considera apenas a distribuição dos atratores e bacias estocásticas, e outra que discretiza as variedades estáveis e instáveis. O primeiro método é aplicado inicialmente aos osciladores de Helmholtz e Duffing sob excitação harmônica ou paramétrica com parâmetros incertos ou ruído adicionado ao carregamento determinístico. Eles demonstram as capacidades adaptativas dos métodos propostos, aumentando a qualidade sem aumentar demasiadamente o custo computacional. A dependência do tempo das respostas estocásticas é demonstrada, com longos transientes influenciando o comportamento global. Por fim, discute-se o efeito das incertezas e ruídos nas áreas das bacias, distribuições de atratores e limites das bacias, que podem ser usados para avaliar a integridade dinâmica de sistemas com bacias coexistentes. Em seguida, dois Sistemas Micro-Eletro-Mecânicos (MEMS) atuados eletricamente, uma microviga em balanço e um microarco imperfeitos, são investigados. O efeito do ruído adicionado e da incerteza paramétrica em ambas as estruturas é demonstrado. Os resultados destacam a importância da aleatoriedade na dinâmica global de sistemas dinâmicos com atratores coexistentes. / [en] The aim of this thesis is to provide tools for the global analysis of nondeterministic dynamical systems with competing attractors considering parameter uncertainty and noise and apply them to real-world engineering problems. For this, an adaptative phase-space discretization strategy based on the classical Ulam method is proposed. Initially, a review of the mathematical definitions of dynamical systems, parametric uncertainty, and noise is presented, and the effect of randomness on the global dynamical structures is highlighted. Discretized transfer operators with the necessary modifications due to parameter uncertainty are derived. The stochastic basin of attraction and attractors’ distributions replace the usual basin and attractor concept. For parameter uncertainty cases, the phase-space is augmented with the corresponding probability space, resulting in a collection of transfer operators for which mean results are obtained. Two adaptative phase-space discretization strategies are proposed, one which only considers the attractors’ distribution and stochastic basins, and another that discretizes the stable and unstable manifolds. The first method is initially applied to the Helmholtz and Duffing oscillators under harmonic or parametric excitation with uncertain parameters or added load noise. They demonstrate the adaptive capabilities of the proposed methods, increasing the quality without overly increasing the computational cost. The time-dependency of stochastic responses is demonstrated, with long-transients influencing the global behavior. Finally, the effect of uncertainties and noise on the basins areas, attractors distributions, and basin boundaries are discussed, which can be used to evaluate the dynamic integrity of the competing basins. Then, two electrically actuated Microelectromechanical Systems (MEMS), an imperfect microcantilever and microarch, are investigated. The effect of added noise and parametric uncertainty on both structures is demonstrated. The results highlight the importance of randomness on the global dynamics of dynamical systems with competing attractors.

Das Klimasystem der Erde besteht aus zahlreichen interagierenden Teilsystemen, die sich über verschiedene Zeitskalen hinweg verändern, was zu einer äußerst komplizierten räumlich-zeitlichen Klimavariabilität führt. Das Verständnis von Prozessen, die auf verschiedenen räumlichen und zeitlichen Skalen ablaufen, ist ein entscheidender Aspekt bei der numerischen Wettervorhersage. Die Variabilität des Klimas, ein sich selbst konstituierendes System, scheint in Mustern auf großen Skalen organisiert zu sein. Die Verwendung von Klimanetzwerken hat sich als erfolgreicher Ansatz für die Erkennung der räumlichen Ausbreitung dieser großräumigen Muster in der Variabilität des Klimasystems erwiesen. In dieser Arbeit wird mit Hilfe von Klimanetzwerken gezeigt, dass die Klimavariabilität nicht nur auf größeren Skalen (Asiatischer Sommermonsun, El Niño/Southern Oscillation), sondern auch auf kleineren Skalen, z.B. auf Wetterzeitskalen, in Mustern organisiert ist. Dies findet Anwendung bei der Erkennung einzelner tropischer Wirbelstürme, bei der Charakterisierung binärer Wirbelsturm-Interaktionen, die zu einer vollständigen Verschmelzung führen, und bei der Untersuchung der intrasaisonalen und interannuellen Variabilität des Asiatischen Sommermonsuns. Schließlich wird die Anwendbarkeit von Klimanetzwerken zur Analyse von Vorhersagefehlern demonstriert, was für die Verbesserung von Vorhersagen von immenser Bedeutung ist. Da korrelierte Fehler durch vorhersagbare Beziehungen zwischen Fehlern verschiedener Regionen aufgrund von zugrunde liegenden systematischen oder zufälligen Prozessen auftreten können, wird gezeigt, dass Fehler-Netzwerke helfen können, die räumlich kohärenten Strukturen von Vorhersagefehlern zu untersuchen. Die Analyse der Fehler-Netzwerk-Topologie von Klimavariablen liefert ein erstes Verständnis der vorherrschenden Fehlerquelle und veranschaulicht das Potenzial von Klimanetzwerken als vielversprechendes Diagnoseinstrument zur Untersuchung von Fehlerkorrelationen. / The Earth’s climate system consists of numerous interacting subsystems varying over a multitude of time scales giving rise to highly complicated spatio-temporal climate variability. Understanding processes occurring at different scales, both spatial and temporal, has been a very crucial problem in numerical weather prediction. The variability of climate, a self-constituting system, appears to be organized in patterns on large scales. The climate networks approach has been very successful in detecting the spatial propagation of these large scale patterns of variability in the climate system. In this thesis, it is demonstrated using climate network approach that climate variability is organized in patterns not only at larger scales (Asian Summer Monsoon, El Niño-Southern Oscillation) but also at shorter scales, e.g., weather time scales. This finds application in detecting individual tropical cyclones, characterizing binary cyclone interaction leading to a complete merger, and studying the intraseasonal and interannual variability of the Asian Summer Monsoon. Finally, the applicability of the climate network framework to understand forecast error properties is demonstrated, which is crucial for improvement of forecasts. As correlated errors can arise due to the presence of a predictable relationship between errors of different regions because of some underlying systematic or random process, it is shown that error networks can help to analyze the spatially coherent structures of forecast errors. The analysis of the error network topology of a climate variable provides a preliminary understanding of the dominant source of error, which shows the potential of climate networks as a very promising diagnostic tool to study error correlations.