Atmosphärisches Aerosol kann den Strahlungstransport signifikant beeinflussen. Mineralstaub, der über der Sahara und anderen Wüsten in die Atmosphäre gelangt, ist das hinsichtlich der in letzterer dauerhaft verbleibenden Masse bedeutendste Aerosol. Darüber hinaus sind Saharamineralstaubpartikel nichtkugelförmig, und die Wirkungen dieser Partikeleigenschaft auf den Strahlungstransport in der Erdatmosphäre sind bislang nur ungenügend untersucht worden. Es werden die optischen Eigenschaften, Strahlungs- und Erwärmungseffekte von Saharamineralstaub unter Berücksichtigung der Nichtkugelförmigkeit seiner Partikel quantitativ untersucht, wobei der gesamte, im Hinblick auf den Strahlungshaushalt energetisch relevante Spektralbereich zugrunde gelegt wird. Zunächst werden auf Basis in-situ-gemessener Experimentaldaten die atmosphärischen Umgebungsbedingungen, Größenverteilungen, Brechungsindizes, Bodenalbedo und Partikelgestalt festgelegt, die in einem zweiten Schritt in ein Strahlungstransportmodell einfließen. Mit dessen Hilfe wird in umfangreichen numerischen Simulationen des Strahlungstransports in einer realistischen mineralstaubhaltigen Modellatmosphäre im Vergleich zu Messdaten beispielsweise geklärt, welche Partikelform und Größenäquivalenz angenommener sphäroidaler Modellpartikel am meisten realistisch sind. Des Weiteren werden im Zusammenhang mit der Partikelnichtkugelförmigkeit Sensitivitätsstudien zur Beantwortung der Fragen durchgeführt, inwieweit diese das Strahlungsfeld beeinflusst und zu veränderten Strahlungserwärmungswirkungen führt.

In this thesis, the parallel performance of the finite element-spherical harmonics (FE-PN) method implemented in the general-purpose radiation transport code EVENT is studied both analytically and empirically. EVENT solves the coupled set of space-angle discretized FE-PN equations using a parallel block-Jacobi domain decomposition method. As part of the analytical study, the thesis presents complexity results for EVENT when solving for a 3D criticality benchmark radiation transport problem in parallel. The empirical analysis is concerned with the impact of the main algorithmic factors affecting performance. Firstly, EVENT supports two solution strategies, namely MOD (Moments Over Domains) and DOM (Domains Over Moments), to solve the transport equation in parallel. The two strategies differ in the way they solve the multi-level space-angle coupled systems of equations. The thesis presents empirical evidence of which of the two solution strategies is more efficient. Secondly, different preconditioners are used in the Preconditioned Conjugate Gradient (PCG) inside EVENT. Performance of EVENT is compared when using three preconditioners, namely diagonal, SSOR(Symmetric Successive Over-Relaxation) and ILU. The other two factors, angular and spatial resolutions of the problem affect both the performance and precision of EVENT. The thesis presents comparative results on EVENTs performance as these two resolutions are increased. From the empirical performance study of EVENT, a bottleneck is identified that limits the improvement in performance as number of processors used by EVENT is increased. In some experiments, it is observed that uneven assignment of computational load among processors causes a significant portion of the total time being spent in synchronization among processors. The thesis presents two indicators that identify when such inefficiency occur; and in such a case, a load rebalancing strategy is applied that computes a new partition of the problem so that each partition corresponds to equal amount of computational load.

A general finite difference discretization of the time-dependent radiation transport equation is developed around the framework of an existing steady-state three dimensional radiation transport solver based on the slice-balance approach. Three related algorithms are outlined within the general finite difference scheme: an explicit, an implicit, and a semi-implicit approach. The three algorithms are analyzed with respect to the discretizations of each element of the phase space in the transport solver. The explicit method, despite its small computational cost per time step, is found to be unsuitable for many purposes due to its inability to accurately handle rapidly varying solutions. The semi-implicit method is shown to produce results nearly as reliable as the fully implicit solver, while requiring significantly less computational effort.

The first part of this study is focused on numerical studies of light scattering from a single microscopic particle using the Discrete Dipole Approximation (DDA) method. The conventional DDA formalism is generalized to two cases: (a) inelastic light scattering from a dielectric particle and (b) light scattering from a particle with magnetic permeability u /= 1. The first generalization is applied to simulations of Raman scattering from bioaerosol particles, and the second generalization is applied to confi rmation of irregular invisibility cloaks made from metamaterials. In the second part, radiative transfer in a coupled atmosphere-ocean system is solved to study the asymptotic nature of the polarized light in deep oceans. The rate at which the radiance and the polarization approach their asymptotic forms in an ideal homogeneous water body are studied. Effects of the single scattering albedo and the volume scattering function are studied. A more realistic water body with vertical pro files for oceanic optical properties determined by a Case 1 water model is then assumed to study the e ffects of wavelength, Raman scattering, and surface waves. Simulated Raman scattering patterns computed from the generalized DDA formalism are found to be sensitive to the distribution of Raman active molecules in the host particle. Therefore one can infer how the Raman active molecules are distributed from a measured Raman scattering pattern. Material properties of invisibility cloaks with a few irregular geometries are given, and field distributions in the vicinity of the cloaked particles computed from the generalized DDA formalism con rm that the designated material properties lead to invisibility. The radiative transfer model calculation in deep oceans suggest that the underwater radiance approaches its asymptotic form more quickly than the polarization does. Therefore, a vector radiative transfer solution is necessary for asymptotic light field studies. For a typical homogeneous water body whose scattering property is characterized by the Petzold phase function, a single scattering albedo of w0 > 0:8 is required in order that the asymptotic regime can be reached before there are too few photons to be detected.

The objective of this thesis is to evaluate the sources of the differences between the NO2 satellite retrieval products provided by the Royal Dutch Meteorological Institute (KNMI) and the National Aeronautics and Space Administration (NASA). Ground studies have shown that although both products use the same satellite, these products yield different observations for NO2 tropospheric columns concentrations. This study does not validate either retrieval product, but rather indentifies the main sources for the discrepancy. There are several parameters which allow successful retrieval of NO2 vertical columns. For this study, only the difference between the radiative models and the a priori NO2 chemical transport models were considered relevant. All other parameters, such as cloud properties, slant columns, stratospheric serration and their assumptions, were held constant. Here, the models are referred to by their proprietor's acronym: "TOMRAD" refers to the radiative model used by NASA, "DAK" refers to the radiative model used by KNMI, "TM4" refers to the a priori chemical transport model used by KNMI, and "REAM" refers to the a priori chemical transport model maintained by the School of Earth and Atmospheric Sciences at the Georgia Institute of Technology. Mixing these parameters creates four retrievals for comparison. Many significant differences were identified after comparing these four retrievals. First, there are viewing geometry biases between the port side and the starboard side of the satellite retrieval for each swath. These viewing geometry biases lead to artificial periodicities in the retrievals of NO2 tropospheric vertical columns over a specific coordinate or site, such as a city. Furthermore, there were significant differences found after using different a priori NO2 chemical transport models. The low horizontal resolution of TM4 and the satellite retrieval/TM4 coupling effect compared to REAM leads to considerable questioning of the near real time application of the KNMI NO2 retrieval product. Though the TM4 model performs poorly, TM4 retrievals do perform nearly as well as REAM retrievals at capturing day-to-day variability and the spatial variability of the cities used as examples here. The retrievals using TOMRAD outperformed the retrievals using DAK when compared to the high resolution, hourly REAM a priori chemical transport model. In sum, these findings should lead to better optimizations of both the KNMI and NASA retrievals, and thus make their publicly available data products more reliable and accurate for general use.

Near-field thermal radiation which can exceed blackbody radiation by several orders of magnitude has potential applications in energy conversion devices, nanofabrication, and near-field imaging. The present dissertation provides a comprehensive and thorough investigation of near-field heat transfer between parallel plates at nanometer distances. The first part of this dissertation focuses on the fundamentals of nanoscale thermal radiation through a systematic study on the near-field heat transfer between doped Si plates. In order to calculate the near-field heat transfer, it is important to accurately predict the dielectric function of doped Si. The dielectric function of doped Si which is described by the Drude model is a function of carrier concentration and mobility. Hence, accurate ionization and carrier mobility models for both p- and n-type Si are identified after a careful review of the available literature. The radiative properties calculated using the improved dielectric function agrees to a good extent with measurements performed using a FTIR. The near-field heat transfer between doped Si plates at varying doping levels is then calculated using the improved dielectric functions. Several important and characteristic features of near-field radiation are revealed in the analysis. An interesting issue regarding the maximum achievable nanoscale thermal radiation arises out of the study on near-field heat transfer in doped Si. The second part of this dissertation investigates the maximum achievable near-field thermal radiation between two plates at finite vacuum gaps. Initially, both the emitter and the receiver are assumed to have identical frequency-independent dielectric functions and a cut off in the order of the lattice spacing is set on the upper limit of the wavevector. The energy transfer is maximum when the real part of dielectric function is around -1 due to surface waves. On the other hand, there is a strong relationship between the imaginary part of the dielectric function and the vacuum gap. While the study using frequency independent dielectric function is not realistic, it lays down the guidelines for the parametric optimization of dielectric functions of real materials for achieving maximum near-field heat transfer. A parametric study of the different adjustable parameters in the Drude and Loretz model is performed in order to analyze their effect on the near-field heat transfer. It is seen that the optimized Drude model always results in greater near-field heat transfer compared to the Lorentz model and the maximum achievable near-field heat transfer is nearly 1 order greater than that between real materials. In the third part of this dissertation, the unusual penetration depth and the energy streamlines in near-field thermal radiation are studied. It is seen that unlike far-field radiation, the penetration depth in near-field heat transfer is dependent on the vacuum gap. This unusual feature results in a 10 nm thick SiC film behaving as completely opaque when the vacuum gap is around 10 nm. The energy streamlines inside the emitter, receiver, and the vacuum gap are calculated using fluctuation electrodynamics and errors generated due to thin film optics are pointed out. It is seen that the lateral shift of the streamlines inside the emitter can be greater than that in the vacuum gap for SiC. However, for doped Si, the lateral shift is comparable in the different media. While the study on the penetration depth determines the thickness of the emitter, the streamlines determine the lateral dimension.

The objective of this study was to assess if new cloud datasets, namely horizontal fields of integrated cloud parameters and transects of cloud profiles becoming available from current and future satellites like MODIS and CloudSAT as well as EarthCARE will allow for the reconstruction of three-dimensional cloud fields. Because three-dimensional measured cloud fields do not exist, surrogate cloud fields were used to develop and test reconstruction techniques. In order to answer the question if surrogate cloud fields may represent real cloud fields and to evaluate potential constraints for cloud field reconstruction, statistics of surrogate cloud fields have been compared to statistics of various remote sensing retrievals. It has turned out that except for cloud droplet effective radius, which is too low, other cloud parameters are in line with parameters derived from measurements. The reconstruction approach is divided into two parts. The first one deals with the reconstruction of the cloud fields. Three techniques with varying complexity are presented constraining the reconstruction by measurements to various degrees. Whereas the first one applies only information of a satellite radiometer, the other two constrain the retrieval also by profile information measured within the domain. Comparing the reconstruction quality of the approaches, there is no superior algorithm performing better for all cloud fields. This might be ascribed to liquid water content profiles of the surrogate cloud fields close to their adiabatic reference. Consequently, the assumption of adiabatic liquid water content profiles of the first scheme yields adequate estimates and additional information from profiles does not improve the reconstruction. The second part of the reconstruction approach addresses the reconstruction quality by comparing parameters of radiative transfer describing photon path statistics as well as reflectances. Therefore three-dimensional radiative transfer simulations with a Monte Carlo code were carried out for the surrogate cloud fields as well as for the reconstructed cloud fields. It was assumed that deviations of the parameter simulated for the reconstructed cloud and the surrogate cloud field are smaller when reconstruction is more accurate. For parameter describing photon pathes it has been found that only deviations of geometrical pathlength statistics reflect the reconstruction quality to a certain degree. Deviations of other parameters like photon penetration depth do not allow for either assessing local differences in reconstruction quality by an individual reconstruction scheme or to infer the most appropriate reconstruction scheme. The differences in reflectances do also not enable to evaluate reconstruction quality. They prevent from gaining insight in local accuracy of reconstruction due to effects like horizontal photon transport weakening the relations between microphysical as well as optical properties and reflectances of the column. In order to address these effects, grids of various complexity, derived by applying photon path properties, were used to weight deviations of cloud properties when analyzing the relationships. Unfortunately, there is no increase of explained variance due to the application of the weighting grids. Additionally, the sensitivity of the results to the model set-up, namely the spatial resolution of the cloud fields as well as the simplification and neglection of ancillary parameters, were analyzed. Though one would assume a strengthening of relationships between deviations of cloud parameters and deviations of reflectances due to more reliable sampling and reduced inter-column transport of photons when column size increases, there is no indication for resolutions where an assessment of the reconstruction quality by means of reflectance deviations becomes feasible. It also has been shown that inappropriate treatment of aerosols in the radiative transfer simulation impose an error comparable in magnitude to differences in reflectances due to inaccurate cloud field reconstruction. This is especially the case when clouds are located in the boundary layer of the aerosol model. Consequently, appropriate aerosol models should be applied in the analysis. May be due to the low surface reflection and the high cloud optical depths, the representation of the surface reflection function seems to be of minor importance. Summarizing the results, differences in radiative transfer do not allow for the assessment of cloud field reconstruction quality. In order to accomplish the task of cloud field reconstruction, the reconstruction part could be constrained employing information from additional measurements. Observational geometries enabling to use tomographic methods and the application of additional wavelengths for validation might help, too. / Ziel der Arbeit war die Evaluierung inwieweit Datensätze von Wolkenparametern, horizontale Felder integraler Wolkenparameter und Schnitte vertikal aufgelöster Parameter, zur Rekonstruktion dreidimensionaler Wolkenfelder genutzt werden können. Entsprechende Datensätze sind durch MODIS und CloudSAT erstmals vorhanden und werden zusätzlich mit dem Start von EarthCARE zur Verfügung stehen. Da dreidimensionale Wolkenfelder aus Messungen nicht existieren, wurden zur Entwicklung der Rekonstruktionsmethoden surrogate Wolkenfelder genutzt. Um die Qualität der surrogaten Wolkenfelder abzuschätzen und um mögliche Randbedingungen zur Rekonstruktion aufzuzeigen, wurden Statistiken der surrogaten Wolkenfelder mit denen unterschiedlicher Fernerkundungsprodukte verglichen. Dabei zeigte sich, dass, abgesehen von den gegenüber Messungen zu geringen Effektivradien der Wolkentropfen in den surrogaten Wolkenfeldern, die übrigen Wolkenparameter gut übereinstimmen. Der Rekonstruktionsansatz gliedert sich in zwei Teile. Der erste Teil beinhaltet die Rekonstruktion der Wolkenfelder. Dazu werden drei Techniken unterschiedlicher Komplexität genutzt, wobei die Komplexität durch den Grad der eingebundenen Messungen bestimmt wird. Während die einfachste Technik lediglich Informationen, wie sie aus Messungen mit einem Satellitenradiometer gewonnen werden können, nutzt, binden die anderen Techniken zusätzlich Profilinformationen aus dem beobachteten Gebiet ein. Analysen zeigten, dass keine der Methoden für alle untersuchten Wolkenfelder den anderen Methoden überlegen ist. Dies mag daran liegen, dass die Flüssigwasserprofile der surrogaten Wolkenfelder nur geringfügig von den in der ersten Rekonstruktionsmethode angenommenen adiabatischen Flüssigwasserprofilen abweichen, so dass die Nutzung der Profile kaum zusätzliche Information für die Rekonstruktion liefert. Im zweiten Teil des Rekonstruktionsansatzes wird die Qualität der rekonstruierten Wolkenfelder durch den Vergleich von Parametern des Strahlungstransfers, wie Photonenpfad-Statistiken und Strahlungsgrößen, evaluiert. Dazu wurden sowohl für die surrogaten Wolkenfelder als auch für die rekonstruierten Wolkenfelder dreidimensionale Strahlungstransfersimulationen mit einem Monte-Carlo-Modell durchgeführt. Angenommen wurde hierbei, dass eine bessere Rekonstruktionsqualität durch geringere Abweichungen der betrachteten Strahlungsparameter aus Simulationen mit rekonstruierten und surrogaten Wolkenfeldern gekennzeichnet ist. Bei den Parametern, die die Photonenwege beschreiben, unterstützen lediglich die Abweichungen der geometrischen Photonenweglängen diese These. Weder erlauben die Abweichungen der übrigen Parameter, zum Beispiel der Eindringtiefen, Rückschlüsse auf die lokale Rekonstruktionsqualität der einzelnen Methoden zu ziehen, noch ermöglichen sie die beste Rekonstruktionsmethode zu identifizieren. Auch die Unterschiede der simulierten Reflektanzen können nicht zur Bestimmung der Rekonstruktionsqualität herangezogen werden. Durch Effekte wie horizontale Photonentransporte werden die Zusammenhänge zwischen mikrophysikalischen und optischen Eigenschaften und Reflektanzen der jeweiligen Gittersäule aufgeweicht, und folglich sind keine Rückschlüsse auf die lokale Rekonstruktionsqualität möglich. Um auf entsprechende Effekte einzugehen, wurden für die Analyse Wichtungsfelder unterschiedlicher Komplexität aus Photonenwegeigenschaften generiert, um diese zur Wichtung der Abweichungen der Wolkeneigenschaften zu nutzen. Der Anteil der erklärten Varianz konnte jedoch durch die Nutzung der entsprechenden Wichtungsfelder nicht erhöht werden. Zusätzlich wurden Sensitivitätsstudien hinsichtlich einzelner Vorgaben der Untersuchung durchgeführt. Dazu wurden sowohl der Einfluss der räumlichen Auflösung der Wolkenfelder als auch die Vereinfachung oder Nichtbetrachtung einzelner Modellparameter analysiert. Eine Reduzierung der Auflösung einhergehend mit einem zuverlässigeren Sampling und reduzierten Photonentransport zwischen den Gittersäulen führte zu keinem direkteren Zusammenhang zwischen den Abweichungen der Reflektanzen und den Abweichungen der mikrophysikalischen Eigenschaften. Folglich existiert keine Auflösung, die die Anwendung des Verfahrens ermöglichen würde. Ebenso wurde gezeigt, dass die unzureichende Einbeziehung von Aerosolen bei den Strahlungstransfersimulationen einen Fehler verursachen kann, der in der Größe dem Unterschied der Reflektanzen unzureichender Wolkenfeldrekonstruktionen gleichkommt. Dies ist insbesondere der Fall, wenn die Wolken sich innerhalb der Grenzschicht des Aerosolmodells befinden. Entspechend sollte in solchen Situationen dem verwendeten Aerosolmodell besondere Beachtung geschenkt werden. Hingegen ist der Einfluss des Ansatzes, wie die Bodenreflektion beschrieben wird, eher gering. Dies mag an dem verwendeten Modell mit einer geringen Albedo in Kombination mit optisch dicken Wolken liegen. Zusammenfassend kann festgestellt werden, dass die Unterschiede im Strahlungstransfer nicht zur Abschätzung der Rekonstruktionsqualität der Wolkenfelder herangezogen werden können. Um dem Ziel einer dreidimensionalen Wolkenfeldrekonstruktion näher zu kommen, könnten beim Rekonstruktionsteil Informationen aus zusätzlichen Messungen als Vorgaben genutzt werden. Ebenso könnten Beobachtungsgeometrien, welche die Anwendung tomographischer Methoden erlauben, sowie zusätzliche Wellenlängen zur Validierung der Rekonstruktionsergebnisse verwendet werden.

We have used a combination of hydrodynamical simulations, dust radiative transfer, and an empirically based analytical model for galaxy number densities and merger rates in order to physically model the bright high-redshift submillimeter-selected galaxy (SMG) population. We report the results of three projects: In the first we study the dependence of a galaxy’s observed-frame submillimeter (submm) flux on its physical properties. One of our principal conclusions is that the submm flux scales significantly more weakly with star formation rate for starbursts than for quiescently star-forming galaxies. Consequently, we argue that the SMG population is not exclusively merger-induced starbursts but rather a mix of merger-induced starbursts, early-stage mergers where two quiescently star-forming disk galaxies are blended into one submm source ("galaxy-pair SMGs"), and isolated disk galaxies. In the second work we present testable predictions of this model by demonstrating how quiescently star-forming and starburst SMGs can be distinguished from integrated data alone. Starbursts tend to have higher luminosity, effective dust temperature, global star formation efficiency \((L_{IR}/M_{gas})\), and infrared excess \((L_{IR}/L_{FUV})\) and tend to lie significantly above the star formation rate-stellar mass relation defined by quiescently star-forming galaxies. These diagnostics can be used to observationally determine the relative contribution of quiescently star-forming and starburst galaxies to the SMG population. In the final work we present the SMG number density, cumulative number counts, and redshift distribution predicted by our model. We show that, contrary to previous claims, the observed SMG number counts do not provide evidence for a top-heavy initial mass function. We also show that starbursts and galaxy-pair SMGs both contribute significantly to the bright SMG counts, whereas isolated disks contribute significantly only at the faint end. / Astronomy

In extragalactic astronomy, a central challenge is that we cannot directly watch what happens to galaxies before and after they are observed. This dissertation focuses on linking predictions of galaxy time-evolution directly with observations, evaluating how interactions, mergers, and other processes affect the appearance of elliptical galaxies. The primary approach is to combine hydrodynamical simulations of galaxy formation, including all major components, with dust radiative transfer to predict their observational signatures. The current paradigm implies that a quiescent elliptical emerges following a formative starburst event. These trigger accretion onto the central supermassive black hole (SMBH), which then radiates as an active galactic nucleus (AGN). However, it is not clear the extent to which SMBH growth is fueled by these events nor how important is their energy input at setting the appearance of the remnant. This thesis presents results drawing from three phases in the formation of a typical elliptical: 1) I evaluate how to disentangle AGN from star formation signatures in mid-infrared spectra during a dust-enshrouded starburst, making testable predictions for robustly tracing SMBH growth with the James Webb Space Telescope ; 2) I develop a model for the rate of merger-induced post-starburst galaxies selected from optical spectra, resolving tension between their observed rarity and merger rates from other estimates; and 3) I present results from Hubble Space Telescope imaging of elliptical galaxies in galaxy clusters at 1 < z < 2, the precursors of present-day massive clusters with \(M \sim10^{15}M_{\odot}\), demonstrating that their stars formed over an extended period and ruling out the simplest model for their formation history. These results lend support to a stochastic formation history for ellipticals driven by mergers or interactions. However, significant uncertainties remain in how to evaluate the implications of galaxy appearance, in particular their morphologies across cosmic time. In the final chapter, I outline an approach to build a "mock observatory" from cosmological hydrodynamical simulations, with which observations of all types, including at high spatial and spectral resolutions, can be brought to bear in directly constraining the physics of galaxy formation and evolution. / Astronomy

In this thesis I focus mainly in studying the process of star formation in both high redshift, and local star forming galaxies, by using an observational technique called integral field spectroscopy (IFS). Although these investigations are aimed at studying the star formation properties of these objects, throughout this work I will also discuss the geometric, kinematic, and chemical structures in the inter-stellar medium of these galaxies, which are intimately connected with the process of star formation itself. The studies presented here were conducted under the umbrella of two different projects. First, the HETDEX Pilot Survey for Emission Line Galaxies, where I have studied the properties of Ly-alpha emitting galaxies across the 2<z<4 range, with an emphasis in trying to understand the process by which Ly-alpha photons, produced in large quantities in the active star forming regions, are able to escape the ISM of these objects, allowing us to detect them in the Ly-alpha line. The second project from which results are presented here is the VIRUS-P Exploration of Nearby Galaxies (VENGA), an ongoing campaign to obtain spatially resolved spectroscopy over a broad wavelength range for large portions of the disks of 30 nearby spiral galaxies. In this thesis, the VENGA data is used to study the physical parameters that set the rate of star formation in the different environments present within galaxies in the local universe. / text