Distributed Acoustic Sensing (DAS) converts fiber-optic cables into spatially dense seismic arrays and has emerged as a versatile and cost-effective technology for continuous seismic monitoring. However, the lower signal-to-noise ratio of individual DAS channels, the large data volumes generated by dense channel spacing, and the strong variability between field deployments pose significant challenges for automated signal processing. This thesis addresses these challenges by developing flexible frameworks for modeling, processing, and interpreting DAS data in the context of microseismic monitoring, covering event detection, phase arrival picking, event localization, and aspects of reservoir characterization. For event detection and phase arrival picking, supervised machine learning models are trained on synthetic data generated by two independently implemented physics- based numerical tools: a staggered-grid finite-difference solver for the elastic wave equation, used to generate synthetic DAS seismograms, and a Lax–Friedrichs fast sweeping scheme for the anisotropic eikonal equation, used to compute first-arrival travel times for training data labeling and event localization. Event detection is formulated as a binary semantic image segmentation problem using a lightweight BASNet-based convolutional neural network with multi-scale supervision, while phase arrival picking extends this approach with a UNETR-based vision transformer architecture capable of capturing long-range dependencies in P- and S-wave moveout patterns. Models trained solely on synthetic data demonstrate robust performance in detecting microseismic events and identifying phase arrivals in unseen field recordings under varying conditions. However, synthetic data cannot perfectly reproduce real data distributions, leading to generalization issues caused by the domain gap between simulated and observed waveforms. To bridge this gap, a CycleGAN-based image- to-image translation workflow is introduced that maps synthetic DAS events into the field-data domain while preserving wavefield geometry, enabling the generation of more realistic training data and thereby improving model performance on field recordings. The proposed methods are validated on three field datasets covering geothermal energy extraction at the Utah FORGE site, carbon storage at the CO2CRC Otway Project, and unconventional hydrocarbon production at the Hydraulic Fracturing Test Site II (HFTS-2). Although DAS channels exhibit a lower signal-to-noise ratio than conventional geophones, the FORGE detection model identifies more than three times as many events since fibers can be deployed near the stimulated intervals, whereas the geophone installations are restricted to shallower dedicated monitoring wells located farther from the reservoir. At the CO2CRC Otway Project, two localization approaches are applied and benchmarked against a reference catalog. A waveform-stacking-based approach is found to be highly sensitive to velocity-model errors, causing inter-phase interference and coherent stacking artifacts that lead to unreliable event locations. In contrast, a probabilistic inversion based on picked DAS arrivals produces plausible event locations; however, standard uncertainty quantification turns out to be infeasible for dense DAS arrays due to inter-channel correlations of picks that cause the posterior distribution to collapse. At HFTS-2, the full processing pipeline is applied, including phase arrival picking using CycleGAN domain-translated training data, event localization, and inter- fiber event association. Training the phase picker on domain-translated data yields approximately 10% more high-quality picked events and approximately 11% higher inlier pick accuracy compared with training on purely synthetic data. The resolution of the resulting DAS-based event catalog is restricted to the intermediate region around the fibers and therefore does not resolve the full extent of the reservoir. Consequently, important subsurface features, such as the seismic gap observed at HFTS-2, cannot be recovered from DAS microseismic data alone. The integrated analysis of microseismicity and low-frequency DAS strain data is used to characterize the stimulated reservoir volume and the triggering mechanisms of induced seismicity during hydraulic fracturing at HFTS-2. This joint analysis demonstrates that microseismic event clouds are not always reliable proxies for fracture geometry, as hydraulic fractures are observed to propagate aseismically through stratigraphic layers. The combination of microseismicity and low-frequency DAS strain measurements enables a more refined interpretation of dynamic fracturing processes, andfurtherallowsderivingreservoirpropertiesmoreprecisely. Forexample, the direct measurement of strain rates along a vertical fiber enables the hydraulic diffusivity to be estimated by fitting the back-front equation to parabolic strain patterns. The results of this thesis suggest that DAS should currently be regarded as a complementary technology rather than a replacement for conventional microseismic monitoring systems. If future improvements enhance DAS signal quality to levels comparable to those of downhole geophones, DAS could become a flexible and cost-effective alternative. The methods developed in this thesis provide a practical framework for automated DAS data processing. As DAS technology continues to evolve and its use in geophysical monitoring becomes increasingly widespread, such automated and reproducible processing workflows will be essential to utilize the full scientific and operational value of DAS data.

To fully understand how and why immune function varies, we must study immune responses within the ecological contexts where they naturally evolve and operate. This thesis aimed to bridge the gap between laboratory and field-based studies in immunity. In Chapter 1, we investigated the prevalence and ecological relevance of wounding in wild-caught Drosophila melanogaster, a commonly used infection route to induce systemic infections in immune studies. Our findings revealed that wounding is frequent in natural populations of D. melanogaster and likely represents a significant selective pressure. A total of 31% of individuals showed signs of physical damage, such as missing appendages or melanised wounds. We also found encapsulated parasitoids and ectoparasitic mites, emphasising the diverse sources of injury in the wild. Strikingly, melanised spots were common on the ventral abdomen of females but absent in males, suggesting a link to mating-related injury. Building on this, Chapter 2 focused on copulatory wounding, injuries that result from mating. We found that increased mating frequency led to more severe genital and abdominal wounds, indicating that copulatory wounding may contribute to sexual conflict and may represent an additional cost of polyandry. Together, Chapters 1 and 2 highlight the injury-related pressures flies experience, which may have a greater impact on immune responses in wild populations due to their constant exposure to microbial pressure and environmental variability, such as fluctuations in food availability, temperature, and parasite frequency. To assess whether such ecological differences translate into variation in immune function between lab and wild flies, in Chapter 3, we examined how immune responses vary across lab-adapted and wild populations by comparing virulence (host mortality), resistance, and tolerance in wild-collected, wild lab-acclimated, and long-term lab-adapted flies challenged with Providencia burhodogranariea. Although wild-collected flies exhibited lower survival, core immune traits like resistance and tolerance did not differ between lab and wild flies. Furthermore, we showed that lab acclimatisation to the lab increased survival in the wild lab-acclimatised flies. This indicates that environmental pressures can play a significant role in shaping immune function. We hypothesised that wounding might allow cuticular microbiota and pathogens to enter the body and cause a systemic infection, which could affect pathogen virulence evolution. To test this, in Chapter 4, we turned to the pathogen side of the interaction to assess how cuticular microbiota influence the evolution of virulence. Serial passage experiment showed that P. burhodogranariea evolved higher virulence when it evolved alone than when it co-evolved with common cuticular microbes such as Acetobacter pasteurianus or Lactoplantibacillus plantarum. Microbial co-infection suppressed pathogen exploitation but did not prevent the evolution of per-parasite pathogenicity, suggesting a role for microbial competition in shaping pathogen evolution. Together, this thesis underscores the need to consider host-pathogen interactions as dynamic processes, shaped by ecological context. It highlights how both host and pathogen traits can be affected in response to persistent pressures in natural environments and offers a foundation for future investigations into how ecological pressures shape host-microbe interactions in natural systems.

Colorectal cancer is the third most common and second most deadly cancer worldwide. Over the last decades, a lot of improvements in the survival of early stages of CRC have been made. Therapy regimens have improved over the last decades, elevating the overall survival of CRC patients with stage I-III. However, the majority of patients die from the metastatic burden of the disease. Early stratification of patients highly prone for metastasis employing causal biomarkers with subsequent intervention is essential to aid in the survival of these patients. An emerging key metastatic biomarker is Metastasis-associated in Colon Cancer 1 (MACC1), which is directly linked to reduced survival, metastasis formation and therapy resistance. Its causal nature for metastasis formation makes it an excellent prognostic tool and therapy target. Circulating cell free RNA can be used to stratify patients highly prone to metastasis. Novel intervention strategies targeting MACC1 are still missing. Therefore, the first aim of this project was to characterize the novel small molecule Compound 22 for the transcriptional inhibition of the MACC1 gene. To do so, ADMET, in vivo and cross entity studies were conducted. Compound 22 displayed drug-like characteristics and antimetastatic properties in cell lines and CRC CDX mouse models. Moreover, the mechanism of action was deconvoluted. The NFκB signaling pathway was reduced, leading to a reduction of the MACC1 gene expression. Compound 22 effectively blocked the TNF-α induced activation of the MACC1 gene expression. This has been shown to be facilitated by a reduced activity of p105 and the subsequent reduction of transcription factor activity of this pathway. Furthermore, MACC1 employs other metastatic genes, such as S100A4, to increase the motility of cancer cells. S100A4 can directly interact with proteins of the cytoskeleton, granting cancer cells higher mobility. Moreover, it induces the degradation of the ECM and ultimately leads to higher metastasis formation. The second aim of this project was focused on characterizing repurposed and novel S100A4 transcriptional inhibitors. First, the repurposed compound cantharidin and its analogue were characterized in CRC cancer cell lines. Cantharidin and its analogue reduced the MACC1 and S100A4 gene expression and functions mediated by them. Further, a high-throughput screen was conducted to search for novel S100A4 transcriptional inhibitors. E12 was identified and its antimetastatic effects were demonstrated in CRC cancer cell lines and in a CRC CDX metastatic mouse model. The identification of repurposed and novel antimetastatic compounds targeting MACC1 and S100A4 represent promising intervention strategies for personalized care of patients highly prone to metastasis formation. A combination of MACC1 and S100A4 targeted intervention shows synergistic effects, which can be exploited for improvement of patient survival.

Die fortschreitende Digitalisierung verändert Kommunikations- und Arbeitspraktiken zivilgesellschaftlicher Organisationen und eröffnet neue Möglichkeiten der Partizipation, Sichtbarkeit und Ressourcengewinnung. Zugleich stellen Social Media ressourcenschwache, häufig informell organisierte Vereinigungen vor neue Anforderungen und Spannungsfelder. Vor diesem Hintergrund untersucht die Studie aus neo-institutionalistischer Perspektive, in welchem Maße Social Media als institutionalisierte Handlungslogik zur Erreichung sozialer Wirkungen beiträgt und mit welchen weiteren institutionellen Logiken sie kooperiert oder konkurriert. Untersuchungsgegenstand sind Repair Cafés und offene Werkstätten als zivilgesellschaftliche Reparaturorganisationen. Theoretisch verbindet die Studie Ansätze des Neo-Institutionalismus und der institutionellen Logiken mit organisationskommunikations- und medialisierungstheoretischen Perspektiven. Social Media werden dabei als eigenständige institutionelle Ordnung verstanden, die das organisationale Handeln prägt. Im Rahmen eines Mixed-Methods-Designs mit qualitativer Vorstudie und quantitativer Hauptstudie wurden neben einer Social-Media-Logik Logiken traditioneller Medien, der Community und des bürgerschaftlichen Engagements identifiziert und konzeptualisiert. Die Ergebnisse zeigen, dass die Social-Media-Logik den stärksten positiven Einfluss auf die erreichte soziale Wirkung aufweist. Insbesondere zeitlich und räumlich flexible Kommunikation, entgrenzte Reichweite, Interaktivität, Konnektivität, Selbermacher-Kultur und die Unabhängigkeit von Gatekeepern erweisen sich als relevante Merkmale. Darüber hinaus beeinflussen Mediennutzungskompetenzen, Ressourcenverfügbarkeit und Organisationsstruktur die erreichte Wirkung. Die Studie erweitert den Ansatz institutioneller Logiken um eine medientheoretische Perspektive und zeigt empirisch den institutionellen Charakter von Social Media im Feld zivilgesellschaftlicher Reparaturorganisationen. Damit leistet sie einen Beitrag zur Medialisierungs- und Wirkungsforschung im Dritten Sektor.

As compared to the equilibrium state of matter, very little is known about the non-equilibrium state of matter at ultrafast time scales (attoseconds to femtoseconds). A special role in this emerging field is thus played by ab-initio theory. While this approach has begun to show signs of great promise, it has been plagued by an unavoidable mismatch between experimental and theoretical observables. The only numerically feasible ab-initio theory is the time dependent extension of density function theory (TDDFT), and in this theory the natural observables are charge and spin densities and their corresponding currents. In experiments at ultrafast time scales, however, the only practical way to deduce transient electronic states and spin structures is via spectral information. The primary purpose of this thesis is to bridge this divide between state-of-the art theory and experiment. Our approach to solving this problem combines two flavours of TDDFT: real-time TDDFT (RT-TDDFT) and linear-response TDDFT (LR-TTDFT). The former is employed to solve the fundamental time dependent problem, while the latter serves to, at each time step of interest, determine the MCD spectra. In Chapter 4 we have used this approach to study transient magnetic circular dichroism (tr-MCD) in the extreme ultraviolet spectral range (XUV) in bulk- Co and CoPt, which shows excellent agreement between theoretical and experimental spectral data. Deploying our approach, we also show how these complex spectra can be efficiently decoded revealing a wealth of details of the underlying spin dynamics of complex alloy systems at ultrafast time scales. In particular, we compare our theoretical prediction for the tr-MCD for CoPt with experimental measurements and find excellent agreement at different frequencies. In the second part of the research of this thesis we have explored the question of how to probe quasi-particle excitations within TDDFT. The time dependent electronic structure has, in contrast to its ground state counterpart, no resemblance to the quasi-particle excitations of the system. To circumvent this problem we project the transient state of the system onto the Kohn- Sham ground state, and subsequently integrate over energy to find the charge excited at each crystal momenta k, a quantity we denote Nex(k). As we show in Chapter 5 this object provides, in the case of single layer graphene, a picture of momentum space excitations consistent with the two dimensional transient Electron Momentum Density (2D tr-EMD). This both verifies the Nex(k) as providing a good description of momentum space excitations, as well as demonstrating that 2D tr-EMD, which can be obtained experimentally via Compton tomography, forms an experimental basis for investigating momentum space in such 2D materials. Using both these approaches we probe the phenomena of Landau-Zener-Stückelberg (LZS) interference patterns in the first Brillouin zone (1BZ) of graphene, carefully demonstrating how various laser pulse parameters can be tuned to enhance the patterns. Finally in Chapter 6 we deploy the tool of Nex(k) to explore momentum space excitations in monolayer WSe2. The combination between valley selection rule and the large (and opposite, by time reversal symmetry) spin splitting at each valley leads to selection by helical light of valley-spin excitation: spin-valley locking. Taking this further we show that hybrid pulses of circular and infra-red linear laser light can be used to selectively excite spin throughout the Brillouin zone – taking back control from the valleys that have, to date, dominated the ultrafast physics in this material. Such hybrid pulses combine two canonical optical effects: the Bloch acceleration theorem and the valley selection rule, and employing the Nex(k) we show how this combination of fundamental phenomena leads to the desired full control over momentum space. We thus establish a route from laser light to local control over excitation in reciprocal space, opening the way to the preparation of momenta specified excited states at ultra-fast time scales.

Imaging-based spatial transcriptomics (SRT) is an emerging family of technologies that enable transcript-level molecular profiling at subcellular resolution, unlocking new insights into tissue organization, cell-cell interactions and microenvironmental structure. However, the dominant analytical frameworks in the field often rely on segmentation-based paradigms, often borrowed from established single-cell analysis, which arguably introduces inappropriate assumptions and underutilizes the spatial information present in the data. This thesis contributes to the development of a segmentation-free analysis framework for spatial transcriptomics, proposing three dedicated computational tools that operate directly on transcriptomic signal without the need to aggregate observed signal into cell segments: SSAM-lite addresses the need for accessible, segmentation-free cell typing. Building on the kernel density estimation (KDE) approach of the SSAM framework, SSAM-lite provides a performance-optimized, browser-based implementation that enables users to infer cell types directly from transcript coordinates. The method applies spatial smoothing and marker gene-based cell typing in a user-friendly, interactive interface, facilitating exploratory analysis of imaging-based SRT datasets. Validation on benchmark tissue slices shows that SSAM lite replicates expert-annotated cell type maps with high fidelity while lowering technical barriers to entry. Sonar introduces a lattice-based spatial statistics framework for analyzing co-occurrence relationships between feature categories in pixelized cell type maps. The method leverages convolutional kernels and Fast Fourier Transform (FFT) acceleration to efficiently compute spatial co-occurrence curves across multiple radii. These curves quantify how often two features occur at defined spatial distances, enabling topographic analysis of tissue structure. Application to synthetic and biological datasets demonstrates that Sonar can recover meaningful spatial patterns such as periodic islet spacing, intra-islet cell arrangements, and spatial motifs in complex tissues. A Sonar-based generative algorithm is also presented that re-engineers cell-type maps from empirically determined co-occurrence indicators, useful for illustrating the structural information captured by spatial statistics. The ovrlpy tool identifies vertical signal contamination as critical source of noise in SRT data analysis. It defines a vertical signal integrity (VSI) score that can be used identify spatial doublets and tissue folding artefacts. These artefacts can confound downstream analyses, their removal is shown to enhance segmentation-based workflows, yielding more precise gene expression models and cell type cluster separation. Together, these tools contribute to a segmentation-free framework for spatial transcriptomic analysis, grounded in spatial statistics, scalable computation, and quality control using a noise model dedicated to spatial data. Overall, the thesis advocates for a stronger decoupling of spatial transcriptomics analysis from single-cell inspired conventions and adoption of a dedicated analytic landscape, complete with its own methods, benchmarks, and conceptual models, to fully leverage the unique spatial nature of SRT data.

Skeletal muscle contraction relies on a finely coordinated process of excitation–contraction coupling, where an electrical signal leads to an eventual calcium release from the sarcoplasmic reticulum through the type-1 ryanodine receptor (RyR1). Disruption of RyR1 function is associated with severe human diseases, including malignant hyperthermia and congenital myopathies, underscoring its clinical importance. Despite extensive structural studies on isolated RyR1, it remains unclear how native membrane environments and inter-channel interactions regulate its gating mechanism in muscle cells.

In this thesis, electron cryomicroscopy structures of RyR1 embedded within native sarcoplasmic reticulum membranes under physiological ligand conditions at resolutions of 3.5 – 4.9 Å were determined. These structures have uncovered previously unrecognized structural rearrangements specific to native arrays, wherein RyR1 activation is driven predominantly by a ~7° rotation rather than large domain tilting observed in isolated channels. Direct corner-to-corner contacts between adjacent channels stabilize these rotational transitions, forming an intrinsic mechanical basis for cooperative, synchronized calcium release in muscle, also known as coupled gating. By resolving multiple defined functional states, including Apo, Primed intermediates, the fully-open Activated state, and a Ryanodine-locked state, key ligand-induced conformational changes underlying the channel activation sequence were identified.

Moreover, the results of this thesis provide structural insight into disease-linked RyR1 mutations localized precisely at inter-channel interfaces, offering molecular explanations for the pathological calcium leak observed in malignant hyperthermia and congenital myopathies. These findings fundamentally revise the RyR1 gating model, highlight the biological importance of receptor lattice organization, and offer new avenues for therapeutic targeting of RyR1-related muscle disorders.

This thesis developed a platform for the automated execution, analysis and interpretation of glycosylation reactions along with the structured submission of generated data. Over 1000 individual reactions have been executed in the platform. Leading towards this development, the impact of novel technologies on process intensification and stereoselective analysis was demonstrated. At the outset, flow chemistry as a novel technology was used to develop a fully telescoped, continuous process for synthesizing the ionizable lipid ALC-0315, an integral component in one COVID-19 vaccine formulation. By eliminating intermediate chromatographic purification steps, the process demonstrates the potential for simplification and acceleration of a multi-step synthetic process by flow chemistry. The yield was doubled over comparable batch processes with a projected productivity of 7 mmol/h (Chapter 2). Encouraged by these results, the application of novel technologies, towards construction, investigation and prediction of glycosylation reactions was reviewed. Glycosylations were first reported at the end of the 19th century and numerous strategies have been developed to synthesize defined glycans since. These utilize different protective and leaving groups to link diverse monosaccharides, the complexity arising from various structural factors of the coupling partners. Together with often poorly reported yet critical reaction conditions, such as temperature and procedural details, these earned the field a reputation of poor reproducibility and a steep learning curve for newcomers. Flow chemistry has been reported to facilitate challenging linkages and was used to study both discrete (e.g., solvent, model nucleophiles) and continuous (e.g., temperature, reaction time) parameters affecting glycosylation yield and stereoselectivity. Data from small-scale screening campaigns highlighted temperature as a crucial factor in glycosylations (Chapter 3). To accelerate the development of stereoselective glycosylations and examine the effect of temperature thereupon, I constructed a flow chemistry platform that can perform four reactions per hour. We demonstrated rapid analysis (< 1 min) of stereoselectivity using ion mobility spectrometry on crude reaction mixtures of fully protected glycans for the first time. The study found that while temperature modestly affected stereoselectivity, solvent choice played a more significant role (Chapter 4). For predicting ideal reaction conditions and routes towards desired structures, large amounts of structured, high quality data are needed. Although models trained on literature data can predict yield, stereoselectivity and products to some extent, these models would benefit from better and openly accessible data. The automated and thereby reproducible generation of comprehensive databases for glycosylation reactions, conditions, and results, along with raw analytical data, could address data scarcity and enhance quality (Chapters 3 & 5). In order to construct these datasets and created large amounts of structured data, I developed a platform for the fully automated planning, execution, and analysis of glycosylation reactions based on flow chemistry. As a proof of concept, the fully automated screening of 12 different glycosyl nucleophiles and electrophiles 2 at various temperatures was conducted, resulting in over 720 different reactions from 144 distinct substrate combinations. The data obtained identifies challenging couplings, for example glucosamine electrophiles or 3-hydroxyl glucose nucleophiles, suggests required reaction times below 5 min for most cases, and shows structural influences of donors and acceptors on conversion and selectivity (Chapter 6). I expect that exploitation of the developed platform to systematically study glycosylation chemistry and its application to other suitable fields of chemistry to improve data availability, data quality and understanding of influencing factors for respective chemistries.

This dissertation is an ethnographic analysis of the environmental care practices of mainland Japanese migrants in Miyakojima, Okinawa Prefecture. Through the concept of ecologies of care, this study explores how newcomers experience and perform more-than-human care on the islands, and what their practices do beyond their ecological intentions. Set against the historical background of Okinawa-mainland relations and the rapid tourist and migration growth that Miyako is currently experiencing, this dissertation focuses on migrants who narrate their relocation in terms of healing from contemporary Japan. Motivated by emotions of gratitude and obligation to reciprocate the care they received from Miyako, newcomers with often no prior involvement in environmentalism come to position themselves as caretakers of the islands’ natural world. Based on ethnographic fieldwork that combines digital and non-digital methodologies, I examine how environmental care shapes socio-political relationships in Miyako through mundane practices in everyday life. The dissertation argues that environmental practices are key sites where colonial asymmetries between mainland Japan and Okinawa are renegotiated and reproduced. Among the many environmental pressures Miyako is currently facing, migrants’ concerns center on three particular ecological issues: marine debris, stray animals, and agricultural chemicals. These issues are attributed to local residents and informed by postwar narratives circulating in mainland Japan. Translating these discourses into practices, environmental activities offer mainland migrants a vehicle for identity-making, including what it means to be Okinawan, Japanese, and how these identities relate to one another. In so doing, newcomers narrate the local population as indifferent to nature while positioning themselves as moral authorities. Environmental care on Miyako is thus ambivalent: it connects newcomers affectively to the islands while simultaneously reproducing the asymmetries that set mainland Japan and Okinawa apart. By tracing how genuine ecological concerns become entangled with enduring power relations, this study contributes to debates on Okinawa-Japan relations, urban-to-rural migration in Japan, and more-than-human care theory.

This dissertation examines the governance structures for regional revitalisation in Japan, analysing how the interplay between national, prefectural, and municipal levels shapes rural revitalisation strategies. Employing a comparative multi-site case study methodology, this research explores the roles and interactions among different levels of governance in the national government (Tokyo), Fukuoka and Nagasaki Prefectures, Buzen City, and Hasami Town. It argues that Japan's approach to regional development is shifting towards a collaborative governance model that integrates multiple governmental tiers and stakeholders. Thus, the discussion on Japan’s governance should move beyond the traditional centralisation-decentralisation paradigm and should also be considered within the collaborative governance framework. This study contributes to the literature by showing the evolution of governance models and rural revitalisation strategies in Japan after the enactment of the 2014 Regional Revitalisation Law, and discussing their implications for policy and practice. The research underscores the potential of collaborative governance to enhance regional revitalisation efforts. The conclusion points to the need for further research into the balance of power within these governance models and their long-term impacts on rural development.

The coronavirus disease 2019 emerged in December 2019 and has caused millions of deaths to date. Current treatments include antiviral drugs, antibodies, and anti-inflammatory drugs; however, only a limited number of direct-acting drugs are available on the market. SARS-CoV-2 Mpro is a promising therapeutic target, with nirmatrelvir being the most potent inhibitor. However, due to its unfavorable metabolism, it requires co-administration with ritonavir, a CYP3A4 inhibitor, to enhance bioavailability. The urgent need for new antivirals calls for more efficient drug discovery strategies. Traditional drug development approaches, such as high-throughput screening, are costly and time-consuming. Fragment-based drug discovery offers an alternative for accelerating antiviral development, and protein-templated fragment ligation represents a powerful advancement by combining fragment ligation and detection through bioassays. In the first part of this study, a series of small molecular open-chain α-ketoamides, derived from active isatin derivatives, were synthesized to enhance efficacy, and mitigate the toxicity associated with isatin scaffold. The most potent compound 8, exhibited an IC50 value of 75.6 μM, comparable to the unsubstituted isatin derivative 1, and represents a promising starting point for further optimization. Structure-activity relationship studies suggested introducing a hydrophobic substituent at position 4 to fit into the S4 pocket, though an appropriate linker is still required. Starting from the structure of nirmatrelvir, in-situ Strecker reactions were explored as an alternative approach to generate α-aminonitriles, offering greater structural flexibility. The reactions were systematically analyzed in aqueous buffers under physiological conditions. As the Strecker reaction required strongly alkaline conditions (pH 9-10), which are not compatible with the protease assay, protein-templated fragment ligation could not be conducted for this reaction. Instead, an in-situ screening method was developed that efficiently generated diverse α-aminonitriles from an amine library. The hit compound 26 was successfully resynthesized; however, due to spontaneous readdition, it could not be isolated. Acylation of the free amine group can enhance the stability. Further investigations focused on the acylation of the intermediary Strecker product 23 to synthesize nirmatrelvir in-situ, offering an alternative protein-templated synthetic route. The protein-templated effect was demonstrated through the acylation of compound 23 using pentafluorophenol-activated acid, where enhanced inhibition in the FRET assay and increased product formation in HPLC-QToF-MS confirmed the reaction’s efficiency in the presence of protease. A protein-templated fragment ligation screening method could be construct under these conditions, enabling efficient structural variation. Ultimately, nirmatrelvir and its analogs were synthesized via Strecker reaction, isolated, and evaluated for their inhibitory activity. The covalent binding between nirmatrelvir and Mpro was confirmed by protein MS. These results highlight the potential of in-situ Strecker reactions and protein-templated acylation as an efficient method for the discovery of enzyme inhibitors as demonstrated here for SARS-CoV-2 Mpro, enabling the generation and screening of α-aminonitrile libraries. Further efforts should focus on applying this methodology to alternative drug targets which are likely to be inhibited by α-aminonitriles, including cysteine and serine proteases.

Worldwide epidemics of the flu caused by the Influenza A virus lead to thousands of deaths every year, making medical research on drug discovery and vaccine development indispensable. Accordingly, research on the viral machinery and its inhibition is on-going and remains highly relevant. One Influenza A protein that has been of great research interest is the small proton channel matrix protein 2 (M2), which is responsible for acidifying the viral interior during virus entry. Understanding the molecular mechanism of M2 is, on the one hand, essential for antiviral drug development, since inhibitors of the protein have already been identified. On the other hand, it also positions M2 as a model protein for small channel proteins with similar function and structure found in many human-pathogenic viruses, the family of viroporins. The aim of this thesis is to investigate the proton channel mechanism of M2, specifically its opening motion and the influence of carboxylate protonation on this process, by combining infrared spectroscopic and computational methods. To quantify the opening motion of the channel, a combined surface-enhanced infrared absorption (SEIRA) and density functional theory (DFT) approach was applied. To assess the role of carboxylate protonation, pH-dependent SEIRA experiments were complemented by DFT calculations and molecular dynamics (MD) simulations, revealing mutation-induced structural changes and their effect on the opening mechanism. MD simulations also enabled spectral predictions of M2 at different protonation states, allowing comparison with experimental spectra. The combined SEIRA und DFT approach revealed that M2 opens by 17 ± 2° with a three-step transition in a pH range of pH 8 to around pH 4. Interestingly, carboxylate protonation events were found to contribute to these transitions. Mutations of Asp24 and also Asp44 led to changes in both structure and opening mechanism of the protein, where pKA values of the transition were either shifted or completely absent in the variants. This indicates that the role of these aspartates in the mechanism has been underestimated in previous studies. MD simulations showed that structural heterogeneity might play a role in the mechanism as it is influenced by amino acid protonation and also by Asp24 and Asp44 mutation. Furthermore, MD-derived spectra were compared to experimental spectral features in their wavenumber positions, validating the MD as a complementary approach to DFT-based spectroscopy to study viroporins. Finally, cell-free expression was explored for M2 as a model system to establish a method applicable to structurally similar, human-pathogenic viral proteins, which often require high biosafety levels for in-cell expression. Cell-free expression experiments showed promise in expressing and inserting M2 into membranes, but further condition screening is needed to achieve fully active proteins routinely. The results of this work demonstrate that a combined computational and spectroscopic approach can reveal detailed insights into the opening mechanism of the M2 proton channel. The findings suggest that the opening motion is functionally relevant and potentially less symmetrical than previously assumed. It was also shown that the aspartates, and especially Asp24, play a more significant role in the channel mechanism than previously thought. Understanding this mechanism and establishing methods to study it may pave the way for future medical research on influenza A and other human-pathogenic viruses with similar proton channels.

El objeto de la presente disertación lo constituyen los topónimos implantados o recogidos por las empresas castellanas y portuguesas de navegación por las costas e islas de la porción terrestre actualmente designada como América del Sur, durante el periodo comprendido entre el tercer viaje de Cristóbal Colón en 1498 y la expedición de Fernando de Magallanes en 1520, así como la nomenclatura geográfica consignada en las representaciones cartográficas del subcontinente en mapas y globos producidos en Europa durante el primer cuarto del siglo XVI, hoy conservados. En la historiografía sobre el tema, los topónimos asociados con las primeras exploraciones ibéricas de los litorales americanos suelen ser tratados de forma tangencial, circunstancial o anecdótica, o bien como manifestaciones simbólicas de la apropiación territorial, la imposición cultural y la dominación política ejercidas por las potencias imperiales sobre las tierras “descubiertas” y sus pueblos aborígenes. A su vez, en lo concerniente a los primeros mapas del entonces llamado Nuevo Mundo, la nomenclatura geográfica queda comúnmente subordinada a otros aspectos más notorios y ostensibles, como los contornos y la fisonomía del territorio o las leyendas y las ilustraciones. Ahora bien, cuando los topónimos son objeto de análisis, la atención generalmente se enfoca en determinadas obras cartográficas, regiones o denominaciones específicas, y la reflexión se orienta al discernimiento del origen histórico de ciertos nombres o de sus variantes morfológicas y semánticas. Considerando que en este vasto espectro bibliográfico no se identifica un estudio que se ocupe exclusiva e íntegramente de la primera nomenclatura geográfica poscolombina de la actual América meridional, que además del contexto en el cual surgieron los topónimos observe las dinámicas y las diferentes duraciones que estos experimentaron con posterioridad a su implantación, y que, aparte de símbolos con un significado y un trasfondo cultural, los entienda como referentes primordiales en la construcción social del espacio geográfico y como elementos estructurales de su representación cartográfica, la presente investigación se pregunta sobre el papel que los nombres de lugar desempeñaron tanto en el incipiente proceso de configuración territorial del subcontinente como en la producción de sus mapas más tempranos actualmente conocidos. En el marco de dicho cuestionamiento, a partir de un extenso acervo documental que abarca fuentes textuales de diversa índole procedentes del siglo XVI y obras cartográficas que datan de las primeras tres décadas de la centuria, se han recopilado todos los topónimos y referentes geográficos asociados con las empresas ibéricas de navegación que tuvieron lugar entre 1498 y 1520 a lo largo de las costas suramericanas, desde el Darién en el norte hasta el estrecho de Magallanes en el extremo austral. El amplio conjunto de denominaciones e indicaciones obtenido fue posteriormente objeto de organización bajo diversos criterios para proceder a su análisis comparativo y relacional respecto a datos derivados de otras fuentes primarias, de la geografía y la cartografía contemporáneas del continente y de los aportes de otros autores consignados en la bibliografía secundaria consultada. Siguiendo este procedimiento, se evidencia que aquellos topónimos derivados de las primeras incursiones ibéricas por los litorales del llamado Nuevo Mundo que lograron establecerse como convenciones permanentes, conformaron una red de referentes geográficos a partir de la cual empezó a configurarse la geografía moderna del continente americano. A su vez, se constata cómo –paradójicamente– aquellos mapas a través de los cuales comenzó a imponerse el nombre América, experimentaron un estancamiento empírico al reproducir por más de dos décadas un mismo patrón cartográfico. Dicho epónimo –concebido a principios del siglo XVI por una sociedad de humanistas asentada en la región fronteriza franco-alemana–, así como el célebre planisferio de 1507 elaborado por Martin Waldseemüller –en el que la controvertida designación apareció por primera vez en un mapa–, han sido elevados en la literatura especializada –muy por encima de todo el resto de denominaciones y de obras cartográficas surgidas en la época– como los principales protagonistas de lo que se ha dado en llamar la “invención”, la “precognición” o el “descubrimiento mental” de América. Sin embargo, ni el polémico epónimo ni la representación cartográfica en la que originalmente se introdujo, tuvieron nada que ver con el proceso empírico de reconocimiento de las costas de la llamada Tierra Firme de las Indias Occidentales o Nuevo Mundo. Por el contrario, fueron aquellos topónimos –más numerosos aunque menos notorios– resultantes de las primeras incursiones españolas y portuguesas por los litorales de la actual Suramérica, los verdaderos cimientos sobre los cuales se empezó a construir tanto el espacio geográfico continental como el conocimiento y el mapa del mismo.

Augustus the Strong (1670–1733), the Elector of Saxony and King of Poland, stands out in history as a dedicated patron of the arts who transformed Dresden into a pivotal cultural centre in Europe during his reign. Recognising the potential of art to symbolise power and prestige, the Elector constantly expanded his collection by acquiring paintings, sculptures, books, and other works of art. His reign witnessed the reinvention of porcelain outside Asia, with the Meissen manufactory introducing the first European hard-paste porcelain to the market in 1710.

However, Augustus the Strong initially focused on collecting East Asian porcelain, imported by the Dutch East India Company and dispersed through the Netherlands. Between 1699 and 1733, the Elector amassed an impressive Collection, comprising approximately 25,000 pieces of Japanese and Chinese porcelain by 1727. His plans involved adorning an entire palace, the Japanese Palace, exclusively with East Asian and Meissen wares. To achieve this monumental project, vast quantities of ceramics were procured and transported to Dresden through various means. One noteworthy figure in the Elector’s pursuit of porcelain was the Italian diplomat Robert Taparelli, Count of Lagnasco (1659–1735). Acting on Augustus’ behalf, Lagnasco travelled to Amsterdam to secure porcelain purchases. His efforts are meticulously documented through contemporary letters, acquisition lists, specifications, and invoices. The Palace Inventory, compiled from 1721 onwards and extended until at least 1727, further sheds light on the influx of East Asian porcelain through wholesalers, retailers, and the acquisition of entire collections in the 1720s.

Analysing the approximately 25,000 East Asian objects in the Augustean collection by 1727, this research identifies three distinct phases in its development. In the phase between 1699 and 1714, porcelain played only a minor role in the Elector’s acquisition activities. However, a significant surge in East Asian porcelain acquisitions occurred in the phase between 1715 and 1720, marked by concerted efforts to acquire unique pieces from the import centres in the Netherlands. The phase from 1721 until about 1727 witnessed a shift to local Saxon merchants, intensifying the collection’s growth. Notably, this period accounted for 41% of all East Asian porcelain acquisitions in just six years, with private wholesalers and retailers playing a pivotal role.

The dissertation examines the development of Augustus the Strong’s collection of East Asian porcelain between 1699 and 1733 on the basis of contemporary documents and analyses the importance of the private porcelain trade for the growth of the collection, especially in the third decade of the eighteenth century.

Attosecond pulses generated through mhigh-harmonic generation (HHG) have been instrumental in studying ultrafast dynamics on the shortest timescales over the past few decades.

However, the HHG process is inherently inefficient, and conventional XUV optics cause further signal loss. As a result, most attosecond spectroscopy experiments to date have relied on attosecond XUV pulses combined with near-infrared fields, limiting temporal and spectral flexibility.

This PhD work demonstrates the first HHG-based all-attosecond transient absorption spectroscopy experiments. The work also introduces the first plasma-based lens for focusing XUV attosecond pulses.

Sphingosinkinasen (SphKs) katalysieren die Bildung der bioaktiven Lipide Dihydrosphingosin 1-phosphat (dhS1P) und Sphingosin 1-phosphat (S1P) und sind an verschiedenen Krankheitsprozessen wie Krebs und Infektionen beteiligt. Infolgedessen wurden zahlreiche niedermolekulare SphK-Inhibitoren entwickelt, die häufig in experimentellen Studien eingesetzt werden. Ihre zellulären Effekte auf das Sphingolipidom sind jedoch bislang nur unzureichend charakterisiert. In dieser Arbeit wurde der Einfluss von sieben häufig verwendeten SphK-Inhibitoren (5c, ABC294640 (Opaganib), N,N-Dimethylsphingosin, K145, PF-543, SLM6031434 und SKI-II) auf das Sphingolipid (SL)-Profil in verschiedenen Zelllinien untersucht. Für die meisten Verbindungen wurde wie erwartet eine Reduktion der intrazellulären (dh)S1P-Spiegel beobachtet, mit Ausnahme von 5c, das kaum Effekte zeigte. Im Gegensatz dazu führten die SphK2-spezifischen Inhibitoren K145 und ABC294640 in allen Zelllinien zu einer dosisabhängigen, ausgeprägten Erhöhung von dhS1P und S1P. Kompensatorische Effekte über SphK1 konnten ausgeschlossen werden, da vergleichbare Ergebnisse auch in SphK1-defizienten HK-2-Zellen erzielt wurden. Beide Substanzen zeigten zudem eine vernachlässigbare inhibitorische Wirkung auf humane, rekombinante SphKs. Darüber hinaus wurde bei allen getesteten Inhibitoren eine veränderte Aktivität der Dihydroceramid-Desaturase detektiert - ein bereits für ABC294640 und SKI-II bekannter Effekt.

Weitere mechanistische Analysen zeigten, dass sowohl ABC294640 als auch K145 die de novo-Synthese von SL beeinflussen, indem sie die Aktivität der 3-Ketodihydrosphingosin-Reduktase erhöhen und gleichzeitig die Dihydroceramid-Desaturase hemmen. Zusätzlich führte ABC294640 zur Aktivierung des Salvage-Pathways, was eine verstärkte Rückführung von Sphingosin begünstigte. Zusammen mit einer unveränderter SphK-Aktivität förderte die erhöhte Substratverfügbarkeit die Bildung von überschüssigem S1P. In ABC294640-behandelten Zellen trug zudem eine Hemmung des S1P-Abbaus über die S1P-Lyase zur Akkumulation bei, während K145 den Übergang von S1P in nachgeschaltete Abbauprodukte zu begünstigen scheint.

Zusammenfassend zeigte keiner der sieben getesteten SphK-Inhibitoren ein vollständig vorhersehbares oder spezifisches Wirkprofil auf den SL-Stoffwechsel. Die Ergebnisse unterstreichen die Notwendigkeit, zelluläre SL-Profile bei der Verwendung von SphK-Inhibitoren systematisch zu überwachen, und legen nahe, dass bisherige Studien, die SphK Inhibitoren einsetzten, aber keine quantitativen SL-Daten bereitstellten, mit Vorsicht interpretiert werden sollten. Dies gilt insbesondere für ABC294640, das bisher als einziger SphK-Inhibitor in klinischen Studien am Menschen getestet wurde, unter anderem im Rahmen von Studien zu COVID-19, wodurch die hier beobachteten Effekte potenzielle klinische Relevanz erhalten.

Central European forests provide many ecosystem services such as climate regulation, carbon storage, and economic revenue. However, they are increasingly threatened by forest disturbances driven by climate change. Forest disturbances in Central Europe arise from a variety of agents, including logging, windthrow, drought, and bark beetle infestations, which exhibit highly heterogeneous spatial and temporal dynamics. They are often small, subtle, and develop gradually over time. Bark beetle infestations, for instance, are a key forest management challenge, as they need to be detected within a time frame of 10 weeks after infestation to prevent further dispersal, while visible infestation symptoms usually occur later. To support effective forest management, continuous large-scale monitoring is required. A suitable means for that is multispectral satellite remote sensing for its sensitivity to physiological properties of forest canopies. The Sentinel-2 (S2) satellite constellation provides freely available observations with frequent revisits and moderate spatial resolution. This results in dense satellite time series reflecting forest dynamics through temporal patterns of spectral signals. To analyse these patterns, two monitoring paradigms can be distinguished: offline and online monitoring. The former is used to retrospectively detect disturbances for strategic planning, whereas the latter focuses on near-real-time detection of the onset of disturbances for tactical forest management responses. Both monitoring paradigms require the analysis of noisy, irregular, and complex satellite time series data, where disturbance signals are often subtle and partly overlapped by environmental variability. Conventional approaches based on predefined temporal features or statistical thresholding often struggle to capture such complex and nonlinear patterns in a scalable and transferable manner. On the contrary, the growing S2 archives enable Big Data approaches such as Deep Learning (DL), which can learn complex nonlinear temporal patterns directly from large and diverse time series datasets. This allows them to capture subtle and nonlinear disturbance signals that are difficult to model with conventional approaches. Therefore, this thesis aims at advancing the development of DL approaches for offline and online forest disturbance monitoring that are scalable across large areas, with a focus on the detection of small and subtle forest disturbances and the assessment of their detectability limits. The thesis is based on three studies that focus on Germany and Luxembourg as representative Central European countries: The first study presents a DL-based offline forest disturbance monitoring method using irregular S2 time series trained on a large and diverse dataset across heterogeneous forest ecosystems in Central Europe. The model is specifically trained to focus on small disturbances. Through a spatial hold-out validation design, the generalizability to unseen forest ecosystems including different disturbance agents is evaluated. The model achieves state-of-the-art performance even when taking sub-pixel-level disturbances into account. Nevertheless, scalability limits become apparent when ecological conditions change significantly, i.e., under ecological domain shift conditions. The second study develops one of the first DL-based online monitoring methods for reconstruction-based anomaly detection using irregular S2 time series in Central Europe in an operational setting. The model is trained on a large and diverse dataset, partly derived from automated high-resolution disturbance products, enabling scalable data generation. It is evaluated on a dataset with precisely known bark beetle infestation dates under a spatial hold-out validation design, allowing assessment of detection delay under realistic operational conditions. While reliable detections are achieved after approximately 13 weeks, the model fails to consistently detect infestations within the operationally required 10-week detection window, indicating challenges in early detection despite strong overall performance and transferability. The third study benchmarks conventional time-series-based online monitoring methods against the DL-based approach from Study 2 for timely detection of bark beetle infestations. Under realistic operational conditions, including spatial hold-out validation and evaluation of detection date instead of retrospective breakpoint assessment, the DL method clearly outperforms conventional approaches, primarily by reducing false positives. However, all methods fail to reliably detect infestations within the critical early detection window. This indicates limitations that are unlikely to depend solely on the applied algorithms. Taken together, the three studies provide evidence on the capabilities and limitations of DL-based forest disturbance monitoring using S2 time series under realistic large-scale conditions. This thesis advances large-scale forest disturbance monitoring capabilities in Central Europe by developing DL models for offline and online monitoring trained on heterogeneous datasets using irregular S2 time series with robust performance, including production pipelines and – in case of Study 1 – operational deployment. A spatial hold-out validation was consistently carried out to support understanding large-scale application capabilities and potential performance constraints under ecological domain shift in realistic operational settings. While good generalization of the developed DL models was confirmed throughout the studies, a moderate performance decline could be noted in ecological domain shift conditions in Study 1, i.e. when ecological conditions strongly changed, particularly in case of small disturbances. This indicates ecological information constraints when scaling DL models to larger areas. Increasing ecological variability leads to stronger class overlap, making small and subtle disturbances particularly difficult to detect. I conclude that scaling DL models to large areas across heterogeneous ecosystems leads to informational limits of the spectral signal. Furthermore, the inability to detect bark beetle infestations within the critical early detection window throughout modelling paradigms including DL indicates that this limitation is likely not solely methodological, but may reflect fundamental signal constraints of S2 time series with respect to early bark beetle infestation detection in Central Europe. To control the impact of ecological variability on spectral variability, future studies must incorporate ecological anchors such as auxiliary environmental variables as model input to reduce intra-class variability. This would enhance class separability across large areas, leading to improved large-scale applicability of DL methods.

Digital innovation literature explains the scaling of digital ventures through a layered, modular architecture that describes the loose coupling among a venture’s layers, affording design flexibility and near-costless replication. AI-enabled ventures strain these assumptions as they are shifting from deterministic to probabilistic digital cores. Their cores entangle data, models, and service in tight feedback loops; their ability to scale depends on non-scale-free resources such as specialized human expertise and curated data, and their inherent model opacity complicates boundary work with stakeholders. My thesis addresses this conundrum: although tight coupling and opacity would be expected to undermine flexibility and replication, and thus stall AI-enabled ventures, some nevertheless scale rapidly. Synthesizing findings from six studies using qualitative, quantitative, and design science methods, this dissertation advances the theory of scaling ventures with probabilistic cores (TSVPC). The theory’s central claim is that AI-enabled ventures scale not by eliminating the nonscale- free resources but by actively managing them. TSVPC theorizes scaling in AIenabled ventures as a venture-level process driven by a probabilistic core that creates tight contents, model, and service interdependence. It describes the generative mechanism of automation–augmentation calibration, which allows to calibrate the loosening and tightening of layer couplings deliberately over time. External triggers like model drift, entry into novel segments, trust deficits, and non–scale-free constraints induce recalibration. The process is shaped by three levers: an architecture lever that includes a distinct model layer and the architectural cadence of generative coupling, in which inter-layer ties tighten to restore model behavior stability and loosen once stability is regained, yielding a recurring pattern over time; a venture lever that emphasizes templating, MLOps, and productized data work; and a product lever that uses entrepreneurial framing in product boundary negotiations and cultivates complementing explorers within platforms to expand product boundaries when confronted with nearly unlimited complements. These complementing explorers are users who expand boundaries by exploring novel use cases. Together, these levers improve stable model behavior, market acceptance, and the rate at which the organization can manage data, iterate models, and capture value, thereby driving venture growth. The boundary conditions for TSVPC are primarily new and growth-stage ventures, where a probabilistic digital core is central to value creation and where resources such as data, computation, and expertise are significant constraints. Six work packages ground and build TSVPC. The first study (WP1) is a qualitative interview study that introduces non-scale-free constraints in AI-enabled ventures. The second study (WP2) is an interview-based analysis of entrepreneurial framing under opacity, showing how ventures negotiate product boundaries with investors and users and how these outcomes feed back into calibration. The third study (WP3), the single-authored study of the dissertation, follows an action design research approach and specifies resource orchestration for the AI pipeline by sequencing templating, externalized data work, and capability development to absorb drift while preserving replication benefits; it was developed and evaluated in workshops with 16 startups across two accelerators over more than three years. The fourth study (WP4) is a quantitative analysis of large language model platform usage data that identifies inverse generativity on model platforms and the outsized role of complementing explorers in expanding product boundaries and revealing growth plateaus. The fifth study (WP5) is an interview study that continues the work in WP1 and resolves the loose coupling paradox by formalizing a model layer and defining generative coupling as the architectural cadence that integrates with venture-level calibration when operating on a probabilistic digital core. The sixth study (WP6) continues the action design research started in WP3 and provides a sharpened design theory for orchestrating non-scale-free resources in AIenabled ventures, refined through repeated accelerator cycles with founders as end users and practitioners as domain experts. The contribution of the dissertation is threefold. Conceptually, it revises assumptions about digital scaling by articulating an adapted layered modular architecture with a model layer, introducing generative coupling, and integrating resource orchestration and entrepreneurial framing through automation–augmentation calibration. Design knowledge advances through an actionable resource orchestration model that shows how templating, MLOps, and productized data work together to mitigate bottlenecks without sacrificing replication or scope. For practice, TSVPC offers a steering logic for founders and product leaders to govern probabilistic cores, expand product boundaries responsibly, navigate strategic trade-offs between automation and augmentation, and anticipate growth plateaus as resource constraints bind.

Around 90 % of Berlin's heat supply is based on fossil fuels, accounting for more than 40 % of the city's total CO₂ emissions. A central challenge of decarbonization lies in the temporal mismatch between heat supply and demand. Surplus heat is mainly available during the summer months, whereas demand is particularly high during the cold season. High-Temperature Aquifer Thermal Energy Storage (HT-ATES) provides a seasonal storage technology capable of storing large amounts of heat in the subsurface while requiring only little space on the surface. This makes HT-ATES particularly well suited to be a part of the heat transition in urban areas. Several HT-ATES pilot and test projects have been carried out since the 1970s, but most were abandoned – either because suitable heat sources were missing, integration into district heating networks proved uneconomic, or the underlying geochemical processes were not sufficiently understood. Heating the formation water to typically 40–95 °C shifts the natural chemical equilibrium between fluid and aquifer and can trigger partly irreversible reactions that affect both storage operation and environmental impact. These include efficiency losses from scaling and corrosion, permeability losses from pore clogging, and the mobilization of trace elements that may impact water quality. Implementing HT-ATES therefore requires reliable baseline data, obtained through comprehensive geochemical characterization of the reservoir. This is essential for ensuring long-term stable operation, preventing lasting damage to the aquifer, and identifying potential risks at an early stage. The objective of this work is to develop and apply methods for the geochemical characterization of potential HT-ATES formations in the North German Basin (NGB) in order to enable a well-founded risk assessment and early-stage site evaluation. On-site analytics, batch experiments under HT-ATES conditions, and geochemical equilibrium modeling were developed and applied at two Berlin sites: Jurassic sandstones (221–400 m bgl) and the Triassic Schaumkalk (514–558 m bgl). In the methodological part of this work, a closed-loop fluid monitoring system (FluMoMax) was developed for both representative sampling and continuous monitoring during operation. It records pH, redox potential, electrical conductivity, dissolved oxygen, temperature, turbidity, and density at temperatures up to 120 °C and pressure up to 10 bar. In addition, a built-in sighting tube indicates whether scale is forming. This allows the progress of well development and the optimal sampling time to be determined, and changes in fluid chemistry to be tracked in real time. In addition, a handheld XRF spectrometer (hXRF) was calibrated using a raw-data-based modeling approach. The method was successfully applied on-site for porosity determination on sandstone cores (adj. R² = 0.976 dry; R² = 0.82–0.94 moist) and for quantifying scaling-relevant ions (Cl, K, Ca, SO₄, Sr) in saline formation waters of up to 266 g L-1 TDS (R² = 0.80–0.99). FluMoMax and hXRF reduce analytical processing time from days or weeks to minutes or hours and enable geochemical risk assessment directly at the drill site. The hXRF was also used to characterize 200 m of drill core and cuttings from two HT-ATES exploration wells in the Jurassic sandstone. For the five aquifers encountered, an element-based, semi-quantitative assessment was carried out covering reactive phases (carbonates, pyrite, reactive iron), heterogeneity (grain-size proxies Ti, Rb, Al₂O₃), and permeability (Cl). The first Jurassic sandstone below the Rupelian Clay (221–234 m) has the lowest content of reactive phases but is limited to a thickness of 13 m. The fourth aquifer (359–400 m) is favored as the HT-ATES target horizon. Despite its heterogeneity, it combines a substantial thickness with low contents of reactive phases. Anoxic batch experiments were conducted to assess how HT-ATES operation might affect rock–water interactions. Two sandstone samples from the favored Hettangian aquifer were exposed to 0.5 mol L⁻¹ NaCl solution at 30, 45, 60, and 80 °C for reaction times of 1 to 10 days. Three temperature-induced processes were observed: reductive Fe-hydroxide dissolution, endothermic Fe sorption, and incongruent silicate dissolution. Ca, Mg, Si, Sr, Ba, B, and Li accumulated in the fluid with increasing temperature and reaction time, while Fe – and to a lesser extent Mn – decreased markedly after an initial rise, attributed to endothermic sorption onto mineral surfaces. At 80 °C, up to 27% of the total Ca was leached within 10 days into the carbonate-undersaturated fluid, likely originating from structural Ca in Fe-hydroxides, adsorbed Ca, and finely dispersed carbonates. Given the low CaO content of the aquifer material (< 0.1 wt %), this mobilizable pool may gradually become depleted over successive HT-ATES cycles, leaving less dissolved calcium available for carbonate precipitation. The fractured Rüdersdorf Schaumkalk (> 96 % calcite) in Berlin is the first deep carbonate aquifer in the NGB to be hydrochemically characterized for HT-ATES application. The saline, Na-Cl-dominated formation water (130 g/L, pH 6–7, 3 % gas content) is of marine origin and contains both sulfate-reducing bacteria and methanogenic archaea. Geochemical modeling of ten HT-ATES cycles shows pronounced calcite precipitation in the heat exchanger (up to 31 mg/kgw), as well as a significant dissolution potential in reservoir zones with calcite-undersaturated fluid (up to 21 mg/kgw). HT-ATES in siliciclastic sandstones already benefits from initial international operating experience, with research now focusing on water treatment, monitoring, and long-term behavior. For carbonate aquifers, however, fundamental design and operating concepts still need to be developed. The Jurassic sandstones studied here will serve as the storage horizon at the GeoSpeicher Berlin project. Drilling is scheduled to start in 2026, and at a storage temperature of 95 °C, the site will be one of Europe's hottest HT-ATES systems – demonstrating the potential of seasonal heat storage for Berlin's heat transition.

This dissertation examines the political economy of authoritarianism in modern Iran by analyzing how economic conditions, institutional structures, and political selection mechanisms shape political instability, electoral participation, and subnational governance. Across three essays, it studies the relationship between state capacity, social welfare, political behavior, and administrative control during the late Pahlavi monarchy and the Islamic Republic. The first essay investigates Iran’s economic environment between 1960 and 1979. It constructs a Fiscal Capacity of the Government Index and a Household Welfare Index using a Mamdani-type fuzzy inference system with expert-calibrated membership functions. The indices are compared with alternative measures based on principal component analysis, and fuzzy hypothesis testing is used to assess the relationship between fiscal capacity, household welfare, and political instability. The findings indicate that the gap between state fiscal capacity and household welfare widened before the 1979 Revolution. The strongest support is found for the configuration in which government fiscal capacity is high while household welfare remains low. The second essay examines whether economic adversity mobilizes or demobilizes voters in Iran’s electoral-authoritarian system. Using a province–election panel covering six presidential elections between 2001 and 2021, it estimates the relationship between voter turnout and a misery index combining inflation and unemployment. Fixed-effects models, alternative estimators, instrumental-variable analysis, and spatial econometric methods consistently show that greater economic adversity is associated with higher voter turnout. This result supports an interpretation of elections as institutionally constrained arenas for expressing economic grievances. The third essay studies how competence and political loyalty affect the careers of provincial governors. Using an original province–year dataset covering all 31 Iranian provinces from 2000 to 2020, it analyzes gubernatorial retention, rotation, and dismissal through panel-data and instrumental-variable methods. The results suggest that greater competence increases the probability of gubernatorial exit, particularly through rotation, while the effects of political loyalty are weaker and less stable. This pattern is consistent with the strategic redeployment of capable administrators across provinces. Together, the essays show how authoritarian regimes manage tensions between economic performance, welfare provision, political participation, and administrative control.