11723 research outputs found
Sort by
Knowledge Graph Question Answering with Generative Language Models
No full textA Knowledge Graph (KG) is a data structure that stores information about the world in the form of nodes and edges. The nodes represent people, places, things etc., while the edges store the relationships between the nodes. The nodes are also known as entities, while the edges are known as relations or predicates. Several popular search engines today make use of such KGs in the background. Some well-known and freely available KGs are DBpedia and Wikidata.
One way to access information from a KG is through Question Answering. For example, web-based search engines today give people the ability to type their questions and receive answers. Unfortunately, the current state of search engines leaves much to be desired in the complexity of the questions that a user may type. Current search engines work best when the search term is a keyword or a set of words. Processing complete sentences, with complex logical rules, is still an open problem.
One large step in the direction of language understanding has been the arrival of pre-trained Language Models, such as BERT. Such models have been trained on large amounts of text corpus, and surprisingly, some variants of these models, such as T5 and BART, develop a remarkable ability to generate text, the likes of which are difficult to distinguish from that produced by a human author. These models are also called generative Language Models and are a central focus of this thesis.
Given a question by a user, how does one fetch an answer from the KG? This task is commonly known as Knowledge Graph Question Answering (KGQA). One of the techniques is to convert the user's question to a logical form, or a structured query. One popular query language the reader might be familiar with is SQL. SQL, though, is appropriate for relational databases. In the KG world, the analogue would be a language called SPARQL. The task of converting the natural language text to a logical form is known as semantic parsing.
To be able to execute a SPARQL query on a KG, the SPARQL schema must be valid, e.g., it must be syntactically correct, and it should be logically correct, e.g., one can not expect the correct answer if AND is replaced with an OR. The other requirement is that the constants in the query, such as entity and relation IDs, have to be placed in the correct manner in the query.
In this thesis, we explore the abilities of generative Language Models (LMs) in the task of KGQA, with a focus on the semantic parsing approach.
This dissertation hypothesizes that generative LMs can be used effectively for the task of KGQA. We form two research questions over this hypothesis, and to answer these research questions, a series of five topically interconnected publications are presented in a cumulative fashion. We first test two popular generative LMs on the task of semantic parsing. We compare the performance of these models to their non-pre-trained predecessors. To this end, we utilize some well-known and openly available datasets, which address well-known KGs. Later, we develop methods to improve the default performance of such models on this task.
This thesis shows that while generative LMs are not ideal for the semantic parsing task in their default mode, special text-handling mechanisms can be incorporated into the model to improve their performance considerably. With these adaptations, the models produce state-of-the-art performance across five datasets that address four different KGs.
This thesis provides researchers in this field with a set of tools and techniques to work with generative LMs and adapt them appropriately to the task of KGQA. We show in subsequent chapters, that our findings have been used by contemporary researchers to further advance the state-of-the-art. In the end, we produce a new KGQA dataset, built over a smaller domain-specific KG. On this dataset, a challenge was organized in which seven participating teams tried the latest methods, and further pushed the state-of-the-art for this task. All the code and data used and developed in this thesis have been released as open-source
Evaluation zur Lebensqualität von Patienten mit plexiformen Neurofibromen im Rahmen einer Neurofibromatose Typ 1 im Hinblick auf eine Operation der plexiformen Neurofibrome
No full textNeurofibromatose Typ 1 ist eine autosomal-dominant vererbte Erkrankung, die für die Entstehung peripherer Nervenscheidentumoren prädestiniert. Patienten mit dieser Erkrankung entwickeln in der Pubertät häufig Neurofibrome, die operativ entfernt werden können. Es ist jedoch bisher unklar, inwiefern die Operation zu einer Linderung der Beschwerden und einer Verbesserung der Lebensqualität beitragen kann. Das Ziel dieser Studie war es daher, die Lebensqualität von Neurofibromatose-Typ-1-Patienten nach der operativen Behandlung von plexiformen Neurofibromen zu evaluieren.
In die Studie aufgenommen wurden 74 Neurofibromatose-Typ-1-Patienten, die im Zeitraum vom 01.01.2011 bis 31.12.2020 im Universitätsklinikum Hamburg-Eppendorf aufgrund eines plexiformen Neurofibroms operativ behandelt worden waren. Die Datenerhebung erfolgte anhand eines Fragebogens, der vom Studienteam selbst entwickelt wurde und Punkte der allgemeinen Anamnese, NF1 betreffende Punkte und Aspekte der Lebensqualität abfragt.
Die Ergebnisse zeigten, dass die meisten Patienten eine deutliche positive Veränderung der Erkrankung nach der Operation wahrnahmen. Sowohl die soziale Funktionsfähigkeit (Freundschaften, romantische Beziehungen, Stigmatisierung) als auch die physische Funktionsfähigkeit (generalisierte und lokale Schmerzen, Beeinträchtigungen im Alltag) sowie die psychische Gesundheit (Scham, Selbstbewusstsein, Akzeptanz des Körpers, Verstecken der Fibrome, Zukunftsängste) nahmen insgesamt nach der Operation signifikant zu. Mehr als zwei Drittel der Patienten waren mit der Operation zufrieden und würden sie weiterempfehlen.
Die Ergebnisse dieser Studie verdeutlichen, dass eine operative Behandlung von plexiformen Neurofibromen bei Patienten mit Neurofibromatose Typ 1 zu einer Verringerung der Beeinträchtigungen und einer Verbesserung der Lebensqualität beitragen kann. Zukünftige Studien müssen analysieren, welche Patienten weniger zufrieden mit der Operation waren und aus welchem Grund, um das postoperative Ergebnis weiter zu verbessern
Konsil- und interdisziplinärer Visitendienst der Allgemeinmedizin
No full textZiel dieser Mixed-Methods Arbeit war die Evaluation eines einmal wöchentlich stattfindenden interdisziplinären Visitendienstes mit dem Fachbereich Allgemeinmedizin auf einer gefäßchirurgischen Station und der Vergleich mit herkömmlichen Konsilen. Die Ärzte fassten den Visitendienst als hilfreich auf, dies gilt besonders für multimorbide chirurgische Patienten. Die Entdeckung von bisher nicht erkannten medizinischen Problemen, verbesserte medikamentöse Einstellung, sowie die Unterstützung von Ärzten in Weiterbildung waren zentrale positive Aspekte. Als nachteilig wurde die längere Dauer des Visitendienstes gesehen, sowie dass Konsile nach wie vor von Station gestellt werden mussten. Letzteres beruht auf der Tatsache, dass die Ärzte den Kompetenzbereich der Fachärzte für Allgemeinmedizin für komplexe Patienten als unzureichend ansehen und auf einer zu seltenen Frequenz der Visiten. Dass der Kompetenzbereich der Fachärzte für Allgemeinmedizin jedoch ausreichend ist, wurde durch statistische Auswertung zweier quantitativer Datensätze (Patienten des Visitendienstes und von herkömmlichen Konsilen einer Facharztklinik) gezeigt
Influence of collective effects on seeded FEL and mitigation strategies at FLASH
No full textThe Free Electron Laser (FEL) user facility FLASH at DESY is currently being upgraded within the framework of the FLASH2020+ project with the goal of converting one of the beamlines into an externally seeded beamline. In particular, the Echo-Enabled Harmonic Generation (EEHG) method will be used to generate fully
coherent FEL pulses in XUV and soft X-ray wavelength regions down to 4 nm. As one of the most challenging of the planned operation modes, 4 nm EEHG can be especially sensitive to the parasitic beam dynamics, known as collective effects. In this work, we employ analytical models to investigate the influence of Incoherent Synchrotron Radiation and Intra-Beam Scattering on the efficiency of the EEHG setup. We perform numerical simulations to study the effect of Coherent Synchrotron Radiation (CSR) in the seeding section of the beamline on the spectral properties of EEHG bunching. We compare different models of CSR and highlight the importance of such aspects of the models as chamber shielding and transient effects in the parameter space of 4 nm EEHG set-up at FLASH. Furthermore, for CSR, which poses noticeable challenges, we explore a mitigation strategy using short seed laser pulses to optimize the spectral bandwidth. Finally, we demonstrate a promising solution to the Micro-Bunching Instability (MBI) using the recently installed Laser Heater. This thesis is a self-contained detailed investigation of the electron beam dynamics of EEHG in the seeding section in the presence of collective effects and an integral part of ongoing start-to-end simulation efforts for realistic characterization of the performance of the future seeded FEL beamline at FLASH2020
Nitroxyl-Donoren: PKGIα als neue therapeutische Zielstruktur
No full textDie systemische arterielle Hypertonie ist eine weltweit verbreitete kardiovaskuläre Erkrankung, die zu Organschäden und Schlaganfall bzw. Herzinsuffizienz mit Todesfolge führen kann. Trotz vielfältiger Kombinationstherapien wird nach neuen Medikamenten gesucht, die selbst in therapierefraktären Patient*innen effektiv den Blutdruck senken können. Eine vielversprechende therapeutische Alternative zu den verfügbaren Medikamenten, wie Stickstoffmonoxid (NO)-Donoren oder anderen Vasodilatoren, könnte Nitroxyl (HNO) sein. HNO ist die ein-Elektron-reduzierte und protonierte Form von NO, welche eine vasorelaxierende und positive inotrope Wirkung zeigt, jedoch ohne eine Toleranzentwicklung auszulösen. Die HNO-induzierten oxidativen Modifikationen in der cGMP-abhängigen Proteinkinase Iα (PKGIα) und die daraus resultierende Intradisulfidbrückenbildung zwischen den Cysteinen 117 und 195 in der hochaffinen cGMP-Bindedomäne wurden als ein wichtiger Mechanismus für die Regulation der Kinase unter physiologischen und pathophysiologischen Bedingungen postuliert. Der individuelle Beitrag der beiden Cysteine ist jedoch nicht vollständig aufgeklärt. Das Ziel dieser Arbeit bestand darin, die HNO-induzierte Intradisulfidbrückenbildung und den individuellen Beitrag der Cysteine 117 und 195 sowie ihre funktionelle Relevanz zu charakterisieren.
Dies wurde mit verschiedenen Verfahren untersucht: (i) PKGIα-basierte Förster-Resonanz-Energie-Transfer (FRET) Biosensoren für cGMP-Bindung und Konformationsänderung; (ii) In vitro Kinase Assays; (iii) Messungen der VASP Phosphorylierung in isolierten aortalen vaskulären glatten Gefäßmuskelzellen (VSMCs) (iv) Vasorelaxationsmessungen der Arteria femoralis in neu generierten Knockin (KI)-Mausmodellen. Hierzu wurden mittels molekularer Klonierung FRET-Sensoren und mittels CRISPR/Cas9 Technologie drei neue Mauslinien hergestellt, bei denen das Cystein 117 oder 195 bzw. beide gleichzeitig zu nicht oxidierbarem Serin (C117S, C195S, C117/195S) ausgetauscht wurden. Der experimentelle HNO-Donor 1-Nitrosocyclohexylacetat (NCA) und der klinisch relevante Donor CXL-1020 wurden dabei zur Stimulation der Zellen und Gefäße eingesetzt. Beide HNO-Donoren induzierten in FRET-Experimenten eine aktivatorische Oxidation der löslichen Guanylylcyclase (sGC), die durch den prototypischen sGC Inhibitor 1H-(1,2,4)-Oxadiazolo-(4,3-a)-chinoxalin-1-on (ODQ) aufgehoben werden konnte. Die direkte Oxidation durch HNO in Gegenwart von ODQ führte vergleichbar wie die cGMP-Bindung zu einer Konformationsänderung der Sensoren. Dabei waren die FRET-Effekte in allen Mutanten, außer nach NCA Stimulation bei C195S, signifikant reduziert. Obwohl die in vitro Kinase Assays eine gleichermaßen reduzierte PKGIα Aktivität aufgrund der fehlenden Intradisulfidbrückenbildung in allen Mutanten zeigten, waren die CXL-1020 stimulierte VASP Phosphorylierung in aortalen VSMCs und Vasorelaxation der Arteria femoralis isoliert aus den Prkg1-C117S und Prkg1-C195S KI-Mäusen unauffällig. Hingegen zeigten Prkg1-C117/195S Mäuse neben einer ausgeprägten Splenomagalie und einem dramatischen gastrointestinalen Phänotyp eine signifikant verminderte NCA und CXL-1020 induzierte VASP-Phosphorylierung. Beide Cysteine 117 und 195 sind daher unentbehrlich für die oxidativ induzierte und aktivatorische Intradisulfidbrückenbildung, wogegen einzeln mutierte Cysteine durch unterschiedliche Mechanismen kompensiert werden können. Die oxidative Aktivierung ist somit neben der klassischen cGMP-abhängigen Aktivierung ein wichtiger und unverzichtbarer Prozess für die physiologische Funktion der PKGIα in vivo
The Functions of Manuscripts for the Turfan Manichaean Community (9th-11th Centuries)
No full textThis dissertation is a religious-social study of the Manichaean community of the Turfan region during the 9th-11th centuries and its relationship with the Turfan Manichaean manuscripts. Manichaeism, as an official religion of the Qocho Uyghur Kingdom (based in the Turfan region), was replanted to Turfan from the Mongolian Steppe by Uyghurs. It has been known that the Turfan Manichaean community was sustained by Sogdian missionaries and merchants, and Uyghur nobles. Accordingly, the Uyghur Manichaean manuscripts were more often used by the auditors (lay believers), while the Middle Persian, Parthian, and Sogdian Manichaean ones were mainly intended for the elects (priests). Along with the use of the Turfan Manichaean manuscripts, Manichaeism was introduced and practiced in the realm of the Qocho Uyghur Kingdom. The communications between the Turfan Manichaeans and other Central Asian Manichaean communities can be also attested by the Turfan Manichaean materials. But, the previous scholarship has paid too much attention either to the textual contents and linguistic aspects of the Turfan Manichaean manuscripts or to the analysis of historical events in the eastern Manichaean contexts. However, the actual operation of the Turfan Manichaean community and the Turfan Manichaean use of the manuscripts have not yet been sufficiently researched. Therefore, this dissertation investigates how the Mesopotamian dualist religion (Manichaeism) under the Qocho Uyghur administration successfully took root in the Turfan region, and how the Turfan Manichaean community itself was built up with the manifold use of manuscripts. The manuscripts had various functions and significances for the Turfan Manichaean community. This dissertation answers three questions: First, how the texts and pictures on the Manichaean manuscripts functioned in the Turfan Manichaean missionary and religious practices; secondly, how the Manichaean manuscripts that were used by elects functioned to construct the Turfan Manichaean community; thirdly, what roles the Turfan Manichaean manuscripts and texts played regarding the secular sphere and what they meant to the lay Manichaeans. In brief, this dissertation attempts to reconstruct a relatively more complete image of the Turfan Manichaean community by seeing it from the inside, based on the surviving Manichaean textual and pictorial materials from the Turfan region
Extreme Timing Detectors for Particle Identification and Their Physics Application at Future Electron-Positron Colliders
No full textPicosecond timing detectors, e.g. low gain avalanche diodes (LGADs), gained increasing attention in the context of future Higgs factory experiments due to their recent technological advancements. One of the applications of fast-timing detectors in future Higgs factory experiments is time-of-flight (TOF) particle identification (PID). However, the benefits and challenges of incorporating timing technologies into the proposed detector concepts remain unclear. TOF PID is expected to provide efficient PID at low momentum. However, its performance strongly depends on the reconstruction methods and achievable timing resolution of the specific timing technology, which is influenced by other detector requirements. Thus, TOF PID’s influence on the physics program of future experiments is yet to be understood. This study aims to advance our understanding of the role of fast-timing technologies in future Higgs factory experiments.
This study uses physics samples from the central 2020 Monte Carlo production with a full Geant4 simulation of and a corresponding reconstruction in the International Large Detector (ILD). While the study is performed using the ILD as a showcase, the results are also discussed in the context of other proposed future Higgs factory detector concepts.
This study establishes a novel state-of-the-art track length and TOF reconstruction algorithms that significantly outperform previous state-of-the-art used in the ILD. This study discovers the importance of precise track length reconstruction, which has always been neglected due to its negligible contribution, which changes with picosecond timing. Without the novel reconstruction methods, the results of future studies would be unreliable. The momenta spectrum of the majority of charged hadrons in the typical physics processes at the future Higgs factory experiments is shown to be within the reach of the TOF PID. The potential reconstruction and physics applications of the TOF PID stress the importance of using modern reconstruction tools and highlight the potential physics analyses that can benefit from the TOF PID.
TOF PID is a promising technique that can contribute to the physics reach of future Higgs factory experiments. It covers the momentum range containing most of the particles at a typical Higgs factory experiment and nicely complements the blind spots of other high-momentum PID tools.Pikosekunden-Timing-Detektoren, z.B. Low Gain Avalanche Diodes (LGADs), haben aufgrund ihrer jüngsten technologischen Fortschritte im Zusammenhang mit zukünftigen Higgs-Factory-Experimenten zunehmend an Aufmerksamkeit gewonnen. Eine der Anwendungen von Detektoren mit schneller Zeitmessung in zukünftigen Higgs-Fabrik-Experimenten ist die Flugzeit-Teilchenidentifikation (engl. TOF PID). Allerdings sind die Vorteile und Herausforderungen, die sich aus der Einbeziehung von Zeitmessungstechnologien in die vorgeschlagenen Detektorkonzepte ergeben, noch unklar. Es wird erwartet, dass TOF PID effiziente Teilchenidentifikation im Regime kleiner Impulse ermöglicht. Die Leistungsfähigkeit hängt jedoch stark von den Rekonstruktionsmethoden und der erreichbaren Zeitauflösung der spezifischen Zeitmessungstechnologie ab, die wiederum von anderen Detektoranforderungen beeinflusst wird. Daher muss der Einfluss von TOF PID Methoden auf mögliche Messungen an zukünftiger Experimente noch verstanden werden. Diese Studie zielt darauf ab, unser Verständnis der Rolle der Fast-Timing-Technologien in zukünftigen Higgs-Factory-Experimenten zu verbessern. Diese Studie verwendet Daten aus der zentralen 2020 Monte-Carlo-Produktion mit einer vollständigen Geant4-Simulation und einer entsprechenden Rekonstruktion im International Large Detector (ILD). Obwohl die Studie am Beispiel des ILD durchgeführt wird, werden die Ergebnisse auch im Kontext anderer vorgeschlagener zukünftiger Higgs-Factory-Detektorkonzepte diskutiert. In dieser Studie wird ein neuer, hochmoderner Algorithmus zur Rekonstruktion der Spurlänge und der TOF entwickelt, der die bisher im ILD verwendeten Algorithmen deutlich übertrifft. Diese Studie zeigt die Bedeutung der präzisen Rekonstruktion der Spurlänge, die aufgrund ihres geringen Einflusses, der sich mit dem Pikosekunden-Timing deutlich erhöht, immer vernachlässigt wurde. Ohne die neuen Rekonstruktionsmethoden wären die Ergebnisse zukünftiger Studien unzuverlässig. Es wird gezeigt, dass das Impulsspektrum der meisten geladenen Hadronen in den typischen physikalischen Prozessen der zukünftigen Higgs-Fabrik-Experimente innerhalb der Reichweite der TOF PID Methode liegt. Die potenziellen Rekonstruktions- und Physikanwendungen der TOF PID Methode unterstreichen die Bedeutung des Einsatzes moderner Rekonstruktionswerkzeuge und heben die potenziellen Physikanalysen hervor, die von der TOF PID profitieren können. TOF PID ist eine vielversprechende Technik, die zur physikalischen Reichweite der zukünftigen Higgs-Fabrik-Experimente beitragen kann. Sie deckt den Impulsbereich ab, der die Mehrheit der Teilchen in einem typischen Higgs-Fabrik-Experiment enthält, und deckt die Schwächen anderer, für hohe Impulse ausgelegter, PID-Techniken ab
Exploring Nanoparticle Protein Corona Integrity and Drug Distribution from Biodegradable Carriers with Synchrotron-Based Techniques
No full textThe use of nanomaterials in medicine has increased significantly in recent years, particularly due to their unique physical and chemical properties. Their ability to bind various biological molecules, such as proteins, DNA or drugs opens therefore a wide range of possibilities in therapy and diagnostics. One of the biggest challenges in the application of nanoparticles (NPs) is the formation of the so-called protein corona. As soon as NPs are introduced in biological media different biomolecules adsorb to their surface. This biological cover influences the interactions of the nanoparticles with cells and tissues and thus determines their biodistribution,cell uptake and ultimately their efficiency and safety. The integrity of the protein corona can be influenced by various processes; therefore a deeper understanding of these changes is crucial for the development of safe and effective nanomedical applications.
Furthermore, the question of how different layers of the hybrid construct 3 nanoparticle, ligands, proteins 3 can be differentiated and analyzed within these complex environments. By using traditional techniques, it is often challenging to resolve these multi-layered structures that occur in vivo.
This dissertation aims to address these knowledge gaps and advance research in these areas.
Specifically, the work focuses on three projects:
In the first part, the integrity of a labeled pre-formed protein corona around gold nanoparticles is probed inside cells using X-ray fluorescence imaging. This study provides insight into the time dependent degradation in the intracellular environment.
The second part deals with the characterization of gold nanoparticle, labeled ligands and a labeled protein corona using anomalous small angle X-ray scattering. The proof of concept shows that the method is able to distinguish between the three different parts and thus enable a more detailed analysis of the hybrid system.
In the next chapter, biodegradable polyelectrolyte capsules are explored as promising carrier in the field of drug delivery. These carriers can be designed to enhance the bioavailability of hydrophobic drugs and offer controlled or triggered release properties. Understanding the behavior of these carriers within cells is crucial for optimizing their design and effectiveness.
The third project explores how X-ray fluorescence imaging can be used to track the intracellular distribution of a drug delivered by biodegradable microcarriers. This focuses on the release and localization of a Selenium based drug within the cellular environment, providing important data for the development of effective and safe drug delivery systems
Magnetfeldverdrängung in angetriebenem YB2C3O6.48
No full textIn recent years, strongly correlated materials have emerged as promising candidates for hosting next-generation technologies. Due to their rich phase diagrams featuring different competing orders, their macroscopic properties can be controlled on demand with tailored external stimuli. In particular, the development of intense ultrafast light sources in the mid-infrared and terahertz regime enabled the direct and efficient coupling to the low-energy tuning knobs of these quantum solids. Experiments in this novel field have confirmed the generation of transient ferroelectric, magnetic, topological, and superconducting orders.
Cuprate superconductors are the archetypical member of this class of materials, wherein antiferromagnetism, superconductivity, charge and spin order interchange with moderate tuning of the carrier doping or crystal lattice. Remarkably, in underdoped YBa2Cu3O7, signatures of incoherent superconducting fluctuations are present above Tc, extending up to and above room temperature. The most accepted interpretation postulates that superconductivity exists locally in this pseudogap phase, and fluctuations prevent the emergence of long-range order. These observations have sparked a series of experiments aimed at resurrecting the superconducting state by directly driving the interlayer coupling of this layered compound, conjectured as the determining factor for the superconducting pairing. These expectations were positively met by experiments where transient optical properties reminiscent of equilibrium superconductivity were observed in terahertz time-domain spectroscopy.
However, these findings were not conclusive since they were also compatible with a non-superconducting state with enhanced mobility. In order to clarify the nature of the photo-excited state, it was paramount to verify whether it presented the dynamical equivalent of a Meissner effect, considered a fingerprint of the macroscopic coherence characterizing the superconducting order. However, prior to this work, no experimental study had interrogated the magnetic properties of this exotic transient state.
Here, we set out to address this fundamental question. The core of the investigation revolved around the development of the novel Ultrafast Magnetometry experimental technique, combining elements of magneto-optic imaging with the sampling of terahertz pulses. The diamagnetic response of the superconductor was detected by measuring the spatial profile of the magnetic field in its surroundings. Improving on previous designs, the adoption of diamagnetic detectors and advanced analysis techniques made it possible, for the first time, to follow the magnetic dynamics associated with the onset of superconductivity with 〜1 ps time-resolution and 〜1 μT sensitivity.
This experimental technique revealed the presence of a sizable magnetic field expulsion in YBa2Cu3O6.48 upon photo-excitation. Quasi-static simulations confirmed that the sampled magnetic field was caused by a colossal diamagnetic response 〜-0.3, reproduced at equilibrium only by type-II superconductors. This observation is incompatible with a photo-induced increase in mobility. Rather, it underscores the onset of superconducting coherence over macroscopic length scales. Remarkably, the effect was seen to persist up to room temperature and correlated positively with the terahertz optical measurements, suggesting a common physical origin for these two different observables. These experimental findings support the picture of a pseudogap phase in which incipient superconducting correlations are enhanced or synchronized by the drive
Stable isotope ecology of fishes in the Elbe estuary
No full textStable isotope analysis of δ13C and δ15N evolved into a favored tool for ecologists enabling them, among others, to identify aquatic food web structures and nutritional pathways from lower to higher trophic levels. Especially the ecological role of aquatic consumers, such as fishes, can be uncovered as stable isotope analysis provides insights into their feeding strategies, migrations, foraging habits and dietary niche overlaps. Estuaries link marine and freshwater ecosystems, forming isotopic imprints with distinct stable isotope signatures along the salinity gradient and enabling them to be excellent study sites that can be analyzed using stable isotopes (chapter 1).
In the Elbe estuary European smelt Osmerus eperlanus (Linnaeus, 1758) is the dominating fish species that contributes up to 96% of the overall fish abundance, fulfilling a key function. However, the smelt population has declined dramatically in recent years, leading to a necessity to further improve our knowledge on its habitat exploitation in the estuary (chapter 2). Due to smelt ́s anadromous life cycle juveniles and adults share the same areas and resources during springtime. In our study we conducted stomach content and stable isotope analysis of both life stages, providing information on long-term and short-term feeding preferences and isotopic niche overlaps along the salinity gradient. The dietary preferences changed from smaller to larger prey during ontogeny with a cannibalistic feeding habit, whereby feeding presumably depended on food availability. As we observed both overlapped isotopic niches and prey items (mainly mysids and gammarids), the risk of intraspecific competition in smelt can be discussed. However, as the species richness is overall low in the estuary, but biomass rates are high, we assume minor competition pressure within the smelt population. Along the salinity gradient we observed different isotopic niche widths and overlaps, indicating spatial differences in food web structures. Enriched δ15N values in the maximum turbidity zone of the Elbe estuary revealed an extended food chain in this area that mainly affected juvenile smelt. Based on our findings we conclude that juveniles are exposed to the prevailing food supply and conditions, while adults can leave areas when the environmental circumstances are unfavorable.
Various fishes developed different life cycle strategies to exploit estuarine resources, whereby abiotic induced stressors, such as osmoregulation, may hamper their distribution ranges. Ruffe Gymnocephalus cernua (Linnaeus, 1758) and European flounder Platichthys flesus (Linnaeus, 1758) are eponymous for lower regions in temperate European rivers (ruffe-flounder-region). We used stable isotope analysis of δ13C and δ15N of fish data sampled seasonally from spring 2021-2022 to track movements, habitat exploitation, spatial connectivity and potential diet shifts during ontogeny (chapter 3). Our study revealed new insights on fish migrations and habitat use in estuaries. Facultative catadromous flounder use the Elbe estuary especially during early life stages, with juveniles occurring in summer and autumn and migrating upstream of our sampling area. We found two isotopic patterns of flounder throughout the year that indicate two North Sea populations exploiting this area for nursery. The isotope data of ruffe, in contrast, showed residency in the estuary for its entire life cycle, as hardly any marine and riverine isotope signatures were found in the individuals examined. These findings might indicate a distinct brackish water population, while a freshwater population exists further upstream. However, deeper knowledge is needed to confirm our assumption.
To uncover trophic interactions on lower trophic levels on a seasonal scale along the salinity gradient of the Elbe estuary, we investigated δ13C and δ15N of particulate organic matter (POM) and dominant meso- and microzooplankton species (chapter 4). High quality carbon (C) was derived from freshwater sources upstream of the city of Hamburg, whereby C sources further downstream were of lower quality (C:N > 8). Overall, the POM δ13C values reflected a mixture of riverine and coastal C that was used by most taxa studied. Calanoid copepods Eurytemora affinis contained a broader range in δ13C than POM, suggesting that this species fed selectively on POM components. Brackish mysid shrimps of Mesopodopsis slabberi showed the least variation in its isotopic values which did not overlap with dietary niches of other taxa. During winter and autumn zooplankton increased carnivorous foraging (enriched δ15N values), especially in the middle section, due to high suspended matter loads and the limitation of high-quality autochthonous phytoplankton. Zooplankton of the Elbe can therefore deal with rapid changing environmental dynamics through their trophic plasticity (e.g., feeding niches partioning, habitat shifting).
Some aquatic organisms bind biogenic carbon, e.g., in their shells, bones or exoskeletons, that reflect inorganic carbon from the surrounding environment rather than assimilated carbon derived from their diet, which can be problematic while conducting stable isotope analyses. However, the removal of inorganic carbon (IC) is often neglected in aquatic food web studies. Therefore, we analyzed the effects of acidification versus no acid treatment on both δ13C and δ15N ratios of four common estuarine macroinvertebrates (chapter 5). Our results showed that the necessity of acid treatment prior to stable isotope analysis depends on the species-specific carbonate content. As carbonate is especially high in Gammarus spec. and Palaemon longirostris, we here recommend a removal of IC either by acid or mechanical procedures. In contrast, mysid shrimps showed no strong effect in stable isotope values. We further discussed other options to remove IC based on previous studies as a general protocol on how to proceed is missing.
As estuaries are often affected by anthropogenic threats, environmental conditions may change over time, thus shaping fish community compositions, life cycle guilds and abundances. In our 40-year long-term study we compared four seasonal fish monitoring periods from 1980s-2020s (chapter 6). Generally, we observed changes in fish life cycle guilds, especially in marine-estuarine opportunists, anadromous and catadromous species. Overall low fish abundances in the 1980s were potentially induced by poor water quality, due to high concentration of nitrogen compounds, such as NO2 and NH4, that negatively affected fish health and recruitment success. As water quality improved during the 1990s, fish abundances started to increase until 2010. Recently, we observed declining trends in fish stocks again, especially in smelt, flounder, twaite shad and ruffe. Marine-estuarine opportunists, such as herring and whiting, on the other hand increased as they potentially benefit from widening and deepening of the main channel leading to an amplified inflow of saline water. Recent declines in fish abundances were mainly triggered by increased suspended organic matter (SPM) concentrations, reduced river runoff and low oxygen concentrations. We found most significant changes of abiotic parameters and fish stocks in the freshwater and oligohaline areas of the Elbe estuary, where important nursery areas of certain fish species are located.
Movements between habitats can lead to certain stress responses in fish and can further affect fish health (chapter 7). Fish gills are in direct connection with the surrounding water, so the microbial composition of gill mucus reflect environmental conditions and represent a major entry for pathogens. We compared gill microbiota using prokaryotic sequencing of resident ruffe and anadromous smelt in space and time to further establish fish health indicators. To estimate migratory behavior, we integrated a stable isotope approach based on δ13C. Ruffe microbiota composition was mainly influenced by spatial drivers with bacterial indicator taxa Verticiella, Polynucleobacter and Candidatur Megaira, reflecting freshwater residency. Anadromous smelt arriving from the adjacent North Sea in autumn showed a stepwise adaptation of the gill microbiome, whereby Luteolibacter became dominant and hampered bacterial dysbiosis in the host. We found an overall small core biome containing ~30 genera that can cope with estuarine conditions. Hypoxia and high nutrient loads lead to dysbiosis states, largely driven by takeover of opportunistic pathogens (Acinetobacter, Shewanella, Aeromonas).
In conclusion, we revealed new insights into trophic relationships and habitat exploitation of aquatic consumers in the Elbe estuary on a spatial and temporal scale (chapter 8). Based on our study we were able to prove the efficiency of stable isotope analysis in an estuarine ecosystem to track migration patterns and population dynamics of the most important fish species that has not been investigated before. We found four areas along the salinity gradient that are of crucial relevance in understanding the food web of the Elbe estuary. Especially shallow habitats are important for early fish life stages, such as smelt, flounder and ruffe. Therefore, we conclude that our findings could contribute to future conservation strategies in the Elbe estuary