86-XX GEOPHYSICS [See also 76U05, 76V05]
Refine
Year of publication
Document Type
- Doctoral Thesis (29)
- Bachelor Thesis (1)
Is part of the Bibliography
- yes (30)
Keywords
- numerische Modellierung (3)
- Erdbeben (2)
- Subduktion (2)
- inverse theory (2)
- numerical modeling (2)
- subduction (2)
- 3D geomechanical numerical model (1)
- 3D geomechanisch-nummerische Modellierung (1)
- ASPECT (1)
- Abschätzung der Unsicherheiten (1)
Institute
The Greenland Ice Sheet is the second-largest mass of ice on Earth. Being almost 2000 km long, more than 700 km wide, and more than 3 km thick at the summit, it holds enough ice to raise global sea levels by 7m if melted completely. Despite its massive size, it is particularly vulnerable to anthropogenic climate change: temperatures over the Greenland Ice Sheet have increased by more than 2.7◦C in the past 30 years, twice as much as the global mean temperature. Consequently, the ice sheet has been significantly losing mass since the 1980s and the rate of loss has increased sixfold since then. Moreover, it is one of the potential tipping elements of the Earth System, which might undergo irreversible change once a warming threshold is exceeded. This thesis aims at extending the understanding of the resilience of the Greenland Ice Sheet against global warming by analyzing processes and feedbacks relevant to its centennial to multi-millennial stability using ice sheet modeling.
One of these feedbacks, the melt-elevation-feedback is driven by the temperature rise with decreasing altitudes: As the ice sheet melts, its thickness and surface elevation decrease, exposing the ice surface to warmer air and thus increasing the melt rates even further. The glacial isostatic adjustment (GIA) can partly mitigate this melt-elevation feedback as the bedrock lifts in response to an ice load decrease, forming the negative GIA feedback. In my thesis, I show that the interaction between these two competing feedbacks can lead to qualitatively different dynamical responses of the Greenland Ice Sheet to warming – from permanent loss to incomplete recovery, depending on the feedback parameters. My research shows that the interaction of those feedbacks can initiate self-sustained oscillations of the ice volume while the climate forcing remains constant.
Furthermore, the increased surface melt changes the optical properties of the snow or ice surface, e.g. by lowering their albedo, which in turn enhances melt rates – a process known as the melt-albedo feedback. Process-based ice sheet models often neglect this melt-albedo feedback. To close this gap, I implemented a simplified version of the diurnal Energy Balance Model, a computationally efficient approach that can capture the first-order effects of the melt-albedo feedback, into the Parallel Ice Sheet Model (PISM). Using the coupled model, I show in warming experiments that the melt-albedo feedback almost doubles the ice loss until the year 2300 under the low greenhouse gas emission scenario RCP2.6, compared to simulations where the melt-albedo feedback is neglected,
and adds up to 58% additional ice loss under the high emission scenario RCP8.5. Moreover, I find that the melt-albedo feedback dominates the ice loss until 2300, compared to the melt-elevation feedback.
Another process that could influence the resilience of the Greenland Ice Sheet is the warming induced softening of the ice and the resulting increase in flow. In my thesis, I show with PISM how the uncertainty in Glen’s flow law impacts the simulated response to warming. In a flow line setup at fixed climatic mass balance, the uncertainty in flow parameters leads to a range of ice loss comparable to the range caused by different warming levels.
While I focus on fundamental processes, feedbacks, and their interactions in the first three projects of my thesis, I also explore the impact of specific climate scenarios on the sea level rise contribution of the Greenland Ice Sheet. To increase the carbon budget flexibility, some warming scenarios – while still staying within the limits of the Paris Agreement – include a temporal overshoot of global warming. I show that an overshoot by 0.4◦C increases the short-term and long-term ice loss from Greenland by several centimeters. The long-term increase is driven by the warming at high latitudes, which persists even when global warming is reversed. This leads to a substantial long-term commitment of the sea level rise contribution from the Greenland Ice Sheet.
Overall, in my thesis I show that the melt-albedo feedback is most relevant for the ice loss of the Greenland Ice Sheet on centennial timescales. In contrast, the melt-elevation feedback and its interplay with the GIA feedback become increasingly relevant on millennial timescales. All of these influence the resilience of the Greenland Ice Sheet against global warming, in the near future and on the long term.
Enhanced geothermal systems (EGS) are considered a cornerstone of future sustainable energy production. In such systems, high-pressure fluid injections break the rock to provide pathways for water to circulate in and heat up. This approach inherently induces small seismic events that, in rare cases, are felt or can even cause damage. Controlling and reducing the seismic impact of EGS is crucial for a broader public acceptance. To evaluate the applicability of hydraulic fracturing (HF) in EGS and to improve the understanding of fracturing processes and the hydromechanical relation to induced seismicity, six in-situ, meter-scale HF experiments with different injection schemes were performed under controlled conditions in crystalline rock in a depth of 410 m at the Äspö Hard Rock Laboratory (Sweden).
I developed a semi-automated, full-waveform-based detection, classification, and location workflow to extract and characterize the acoustic emission (AE) activity from the continuous recordings of 11 piezoelectric AE sensors. Based on the resulting catalog of 20,000 AEs, with rupture sizes of cm to dm, I mapped and characterized the fracture growth in great detail. The injection using a novel cyclic injection scheme (HF3) had a lower seismic impact than the conventional injections. HF3 induced fewer AEs with a reduced maximum magnitude and significantly larger b-values, implying a decreased number of large events relative to the number of small ones. Furthermore, HF3 showed an increased fracture complexity with multiple fractures or a fracture network. In contrast, the conventional injections developed single, planar fracture zones (Publication 1).
An independent, complementary approach based on a comparison of modeled and observed tilt exploits transient long-period signals recorded at the horizontal components of two broad-band seismometers a few tens of meters apart from the injections. It validated the efficient creation of hydraulic fractures and verified the AE-based fracture geometries. The innovative joint analysis of AEs and tilt signals revealed different phases of the fracturing process, including the (re-)opening, growth, and aftergrowth of fractures, and provided evidence for the reactivation of a preexisting fault in one of the experiments (Publication 2). A newly developed network-based waveform-similarity analysis applied to the massive AE activity supports the latter finding.
To validate whether the reduction of the seismic impact as observed for the cyclic injection schemes during the Äspö mine-scale experiments is transferable to other scales, I additionally calculated energy budgets for injection experiments from previously conducted laboratory tests and from a field application. Across all three scales, the cyclic injections reduce the seismic impact, as depicted by smaller maximum magnitudes, larger b-values, and decreased injection efficiencies (Publication 3).
Plate tectonics describes the movement of rigid plates at the surface of the Earth as well as their complex deformation at three types of plate boundaries: 1) divergent boundaries such as rift zones and mid-ocean ridges, 2) strike-slip boundaries where plates grind past each other, such as the San Andreas Fault, and 3) convergent boundaries that form large mountain ranges like the Andes. The generally narrow deformation zones that bound the plates exhibit complex strain patterns that evolve through time. During this evolution, plate boundary deformation is driven by tectonic forces arising from Earth’s deep interior and from within the lithosphere, but also by surface processes, which erode topographic highs and deposit the resulting sediment into regions of low elevation. Through the combination of these factors, the surface of the Earth evolves in a highly dynamic way with several feedback mechanisms. At divergent boundaries, for example, tensional stresses thin the lithosphere, forcing uplift and subsequent erosion of rift flanks, which creates a sediment source. Meanwhile, the rift center subsides and becomes a topographic low where sediments accumulate. This mass transfer from foot- to hanging wall plays an important role during rifting, as it prolongs the activity of individual normal faults. When rifting continues, continents are eventually split apart, exhuming Earth’s mantle and creating new oceanic crust. Because of the complex interplay between deep tectonic forces that shape plate boundaries and mass redistribution at the Earth’s surface, it is vital to understand feedbacks between the two domains and how they shape our planet.
In this study I aim to provide insight on two primary questions: 1) How do divergent and strike-slip plate boundaries evolve? 2) How is this evolution, on a large temporal scale and a smaller structural scale, affected by the alteration of the surface through erosion and deposition? This is done in three chapters that examine the evolution of divergent and strike-slip plate boundaries using numerical models. Chapter 2 takes a detailed look at the evolution of rift systems using two-dimensional models. Specifically, I extract faults from a range of rift models and correlate them through time to examine how fault networks evolve in space and time. By implementing a two-way coupling between the geodynamic code ASPECT and landscape evolution code FastScape, I investigate how the fault network and rift evolution are influenced by the system’s erosional efficiency, which represents many factors like lithology or climate. In Chapter 3, I examine rift evolution from a three-dimensional perspective. In this chapter I study linkage modes for offset rifts to determine when fast-rotating plate-boundary structures known as continental microplates form. Chapter 4 uses the two-way numerical coupling between tectonics and landscape evolution to investigate how a strike-slip boundary responds to large sediment loads, and whether this is sufficient to form an entirely new type of flexural strike-slip basin.
The Andes are a ~7000 km long N-S trending mountain range developed along the South American western continental margin. Driven by the subduction of the oceanic Nazca plate beneath the continental South American plate, the formation of the northern and central parts of the orogen is a type case for a non-collisional orogeny. In the southern Central Andes (SCA, 29°S-39°S), the oceanic plate changes the subduction angle between 33°S and 35°S from almost horizontal (< 5° dip) in the north to a steeper angle (~30° dip) in the south. This sector of the Andes also displays remarkable along- and across- strike variations of the tectonic deformation patterns. These include a systematic decrease of topographic elevation, of crustal shortening and foreland and orogenic width, as well as an alternation of the foreland deformation style between thick-skinned and thin-skinned recorded along- and across the strike of the subduction zone. Moreover, the SCA are a very seismically active region. The continental plate is characterized by a relatively shallow seismicity (< 30 km depth) which is mainly focussed at the transition from the orogen to the lowland areas of the foreland and the forearc; in contrast, deeper seismicity occurs below the interiors of the northern foreland. Additionally, frequent seismicity is also recorded in the shallow parts of the oceanic plate and in a sector of the flat slab segment between 31°S and 33°S. The observed spatial heterogeneity in tectonic and seismic deformation in the SCA has been attributed to multiple causes, including variations in sediment thickness, the presence of inherited structures and changes in the subduction angle of the oceanic slab. However, there is no study that inquired the relationship between the long-term rheological configuration of the SCA and the spatial deformation patterns. Moreover, the effects of the density and thickness configuration of the continental plate and of variations in the slab dip angle in the rheological state of the lithosphere have been not thoroughly investigated yet. Since rheology depends on composition, pressure and temperature, a detailed characterization of the compositional, structural and thermal fields of the lithosphere is needed. Therefore, by using multiple geophysical approaches and data sources, I constructed the following 3D models of the SCA lithosphere: (i) a seismically-constrained structural and density model that was tested against the gravity field; (ii) a thermal model integrating the conversion of mantle shear-wave velocities to temperature with steady-state conductive calculations in the uppermost lithosphere (< 50 km depth), validated by temperature and heat-flow measurements; and (iii) a rheological model of the long-term lithospheric strength using as input the previously-generated models.
The results of this dissertation indicate that the present-day thermal and rheological fields of the SCA are controlled by different mechanisms at different depths. At shallow depths (< 50 km), the thermomechanical field is modulated by the heterogeneous composition of the continental lithosphere. The overprint of the oceanic slab is detectable where the oceanic plate is shallow (< 85 km depth) and the radiogenic crust is thin, resulting in overall lower temperatures and higher strength compared to regions where the slab is steep and the radiogenic crust is thick. At depths > 50 km, largest temperatures variations occur where the descending slab is detected, which implies that the deep thermal field is mainly affected by the slab dip geometry.
The outcomes of this thesis suggests that long-term thermomechanical state of the lithosphere influences the spatial distribution of seismic deformation. Most of the seismicity within the continental plate occurs above the modelled transition from brittle to ductile conditions. Additionally, there is a spatial correlation between the location of these events and the transition from the mechanically strong domains of the forearc and foreland to the weak domain of the orogen. In contrast, seismicity within the oceanic plate is also detected where long-term ductile conditions are expected. I therefore analysed the possible influence of additional mechanisms triggering these earthquakes, including the compaction of sediments in the subduction interface and dehydration reactions in the slab. To that aim, I carried out a qualitative analysis of the state of hydration in the mantle using the ratio between compressional- and shear-wave velocity (vp/vs ratio) from a previous seismic tomography. The results from this analysis indicate that the majority of the seismicity spatially correlates with hydrated areas of the slab and overlying continental mantle, with the exception of the cluster within the flat slab segment. In this region, earthquakes are likely triggered by flexural processes where the slab changes from a flat to a steep subduction angle.
First-order variations in the observed tectonic patterns also seem to be influenced by the thermomechanical configuration of the lithosphere. The mechanically strong domains of the forearc and foreland, due to their resistance to deformation, display smaller amounts of shortening than the relatively weak orogenic domain. In addition, the structural and thermomechanical characteristics modelled in this dissertation confirm previous analyses from geodynamic models pointing to the control of the observed heterogeneities in the orogen and foreland deformation style. These characteristics include the lithospheric and crustal thickness, the presence of weak sediments and the variations in gravitational potential energy.
Specific conditions occur in the cold and strong northern foreland, which is characterized by active seismicity and thick-skinned structures, although the modelled crustal strength exceeds the typical values of externally-applied tectonic stresses. The additional mechanisms that could explain the strain localization in a region that should resist deformation are: (i) increased tectonic forces coming from the steepening of the slab and (ii) enhanced weakening along inherited structures from pre-Andean deformation events. Finally, the thermomechanical conditions of this sector of the foreland could be a key factor influencing the preservation of the flat subduction angle at these latitudes of the SCA.
The evolution of life on Earth has been driven by disturbances of different types and magnitudes over the 4.6 million years of Earth’s history (Raup, 1994, Alroy, 2008). One example for such disturbances are mass extinctions which are characterized by an exceptional increase in the extinction rate affecting a great number of taxa in a short interval of geologic time (Sepkoski, 1986). During the 541 million years of the Phanerozoic, life on Earth suffered five exceptionally severe mass extinctions named the “Big Five Extinctions”. Many mass extinctions are linked to changes in climate
(Feulner, 2009). Hence, the study of past mass extinctions is not only intriguing, but can also provide insights into the complex nature of the Earth system. This thesis aims at deepening our understanding of the triggers of mass extinctions and how they affected life. To accomplish this, I investigate changes in climate during two of the Big Five extinctions using a coupled climate model.
During the Devonian (419.2–358.9 million years ago) the first vascular plants and vertebrates evolved on land while extinction events occurred in the ocean (Algeo et al., 1995). The causes of these formative changes, their interactions and their links to changes in climate are still poorly understood. Therefore, we explore the sensitivity of the Devonian climate to various boundary conditions using an intermediate-complexity climate model (Brugger et al., 2019). In contrast to Le Hir et al. (2011), we find only a minor biogeophysical effect of changes in vegetation cover due to unrealistically high soil albedo values used in the earlier study. In addition, our results cannot support the strong influence of orbital parameters on the Devonian climate, as simulated with a climate model with a strongly simplified ocean model (De Vleeschouwer et al., 2013, 2014, 2017). We can only reproduce the changes in Devonian climate suggested by proxy data by decreasing atmospheric CO2. Still, finding agreement between the evolution of sea surface temperatures reconstructed from proxy data (Joachimski et al., 2009) and our simulations remains challenging and suggests a lower δ18O ratio of Devonian seawater. Furthermore, our study of the sensitivity of the Devonian climate reveals a prevailing mode of climate variability on a timescale of decades to centuries. The quasi-periodic ocean temperature fluctuations are linked to a physical mechanism of changing sea-ice cover, ocean convection and overturning in high northern latitudes.
In the second study of this thesis (Dahl et al., under review) a new reconstruction of atmospheric CO2 for the Devonian, which is based on CO2-sensitive carbon isotope fractionation in the earliest vascular plant fossils, suggests a much earlier drop of atmo- spheric CO2 concentration than previously reconstructed, followed by nearly constant CO2 concentrations during the Middle and Late Devonian. Our simulations for the Early Devonian with identical boundary conditions as in our Devonian sensitivity study (Brugger et al., 2019), but with a low atmospheric CO2 concentration of 500 ppm, show no direct conflict with available proxy and paleobotanical data and confirm that under the simulated climatic conditions carbon isotope fractionation represents a robust proxy for atmospheric CO2. To explain the earlier CO2 drop we suggest that early forms of vascular land plants have already strongly influenced weathering. This new perspective on the Devonian questions previous ideas about the climatic conditions and earlier explanations for the Devonian mass extinctions.
The second mass extinction investigated in this thesis is the end-Cretaceous mass extinction (66 million years ago) which differs from the Devonian mass extinctions in terms of the processes involved and the timescale on which the extinctions occurred. In the two studies presented here (Brugger et al., 2017, 2021), we model the climatic effects of the Chicxulub impact, one of the proposed causes of the end-Cretaceous extinction, for the first millennium after the impact. The light-dimming effect of stratospheric sulfate aerosols causes severe cooling, with a decrease of global annual mean surface air temperature of at least 26◦C and a recovery to pre-impact temperatures after more than 30 years. The sudden surface cooling of the ocean induces deep convection which brings nutrients from the deep ocean via upwelling to the surface ocean. Using an ocean biogeochemistry model we explore the combined effect of ocean mixing and iron-rich dust originating from the impactor on the marine biosphere. As soon as light levels have recovered, we find a short, but prominent peak in marine net primary productivity. This newly discovered mechanism could result in toxic effects for marine near-surface ecosystems. Comparison of our model results to proxy data (Vellekoop et al., 2014, 2016, Hull et al., 2020) suggests that carbon release from the terrestrial biosphere is required in addition to the carbon dioxide which can be attributed to the target material. Surface ocean acidification caused by the addition of carbon dioxide and sulfur is only moderate. Taken together, the results indicate a significant contribution of the Chicxulub impact to the end-Cretaceous mass extinction by triggering multiple stressors for the Earth system.
Although the sixth extinction we face today is characterized by human intervention in nature, this thesis shows that we can gain many insights into future extinctions from studying past mass extinctions, such as the importance of the rate of change (Rothman, 2017), the interplay of multiple stressors (Gunderson et al., 2016), and changes in the carbon cycle (Rothman, 2017, Tierney et al., 2020).
Centroid moment tensor inversion can provide insight into ongoing tectonic processes and active faults. In the Alpine mountains (central Europe), challenges result from low signal-to-noise ratios of earthquakes with small to moderate magnitudes and complex wave propagation effects through the heterogeneous crustal structure of the mountain belt. In this thesis, I make use of the temporary installation of the dense AlpArray seismic network (AASN) to establish a work flow to study seismic source processes and enhance the knowledge of the Alpine seismicity. The cumulative thesis comprises four publications on the topics of large seismic networks, seismic source processes in the Alps, their link to tectonics and stress field, and the inclusion of small magnitude earthquakes into studies of active faults.
Dealing with hundreds of stations of the dense AASN requires the automated assessment of data and metadata quality. I developed the open source toolbox AutoStatsQ to perform an automated data quality control. Its first application to the AlpArray seismic network has revealed significant errors of amplitude gains and sensor orientations. A second application of the orientation test to the Turkish KOERI network, based on Rayleigh wave polarization, further illustrated the potential in comparison to a P wave polarization method. Taking advantage of the gain and orientation results of the AASN, I tested different inversion settings and input data types to approach the specific challenges of centroid moment tensor (CMT) inversions in the Alps. A comparative study was carried out to define the best fitting procedures.
The application to 4 years of seismicity in the Alps (2016-2019) substantially enhanced the amount of moment tensor solutions in the region. We provide a list of moment tensors solutions down to magnitude Mw 3.1. Spatial patterns of typical focal mechanisms were analyzed in the seismotectonic context, by comparing them to long-term seismicity, historical earthquakes and observations of strain rates. Additionally, we use our MT solutions to investigate stress regimes and orientations along the Alpine chain. Finally, I addressed the challenge of including smaller magnitude events into the study of active faults and source processes. The open-source toolbox Clusty was developed for the clustering of earthquakes based on waveforms recorded across a network of seismic stations. The similarity of waveforms reflects both, the location and the similarity of source mechanisms. Therefore the clustering bears the opportunity to identify earthquakes of similar faulting styles, even when centroid moment tensor inversion is not possible due to low signal-to-noise ratios of surface waves or oversimplified velocity models. The toolbox is described through an application to the Zakynthos 2018 aftershock sequence and I subsequently discuss its potential application to weak earthquakes (Mw<3.1) in the Alps.
Fluids in the Earth's crust can move by creating and flowing through fractures, in a process called `hydraulic fracturing’. The tip-line of such fluid-filled fractures grows at locations where stress is larger than the strength of the rock. Where the tip stress vanishes, the fracture closes and the fluid-front retreats. If stress gradients exist on the fracture's walls, induced by fluid/rock density contrasts or topographic stresses, this results in an asymmetric shape and growth of the fracture, allowing for the contained batch of fluid to propagate through the crust.
The state-of-the-art analytical and numerical methods to simulate fluid-filled fracture propagation are two-dimensional (2D). In this work I extend these to three dimensions (3D). In my analytical method, I approximate the propagating 3D fracture as a penny-shaped crack that is influenced by both an internal pressure and stress gradients. In addition, I develop a numerical method to model propagation where curved fractures can be simulated as a mesh of triangular dislocations, with the displacement of faces computed using the displacement discontinuity method. I devise a rapid technique to approximate stress intensity and use this to calculate the advance of the tip-line. My 3D models can be applied to arbitrary stresses, topographic and crack shapes, whilst retaining short computation times.
I cross-validate my analytical and numerical methods and apply them to various natural and man-made settings, to gain additional insights into the movements of hydraulic fractures such as magmatic dikes and fluid injections in rock. In particular, I calculate the `volumetric tipping point’, which once exceeded allows a fluid-filled fracture to propagate in a `self-sustaining’ manner. I discuss implications this has for hydro-fracturing in industrial operations. I also present two studies combining physical models that define fluid-filled fracture trajectories and Bayesian statistical techniques. In these studies I show that the stress history of the volcanic edifice defines the location of eruptive vents at volcanoes. Retrieval of the ratio between topographic to remote stresses allows for forecasting of probable future vent locations. Finally, I address the mechanics of 3D propagating dykes and sills in volcanic regions. I focus on Sierra Negra volcano in the Gal\'apagos islands, where in 2018, a large sill propagated with an extremely curved trajectory. Using a 3D analysis, I find that shallow horizontal intrusions are highly sensitive to topographic and buoyancy stress gradients, as well as the effects of the free surface.
One third of the world's population lives in areas where earthquakes causing at least slight damage are frequently expected. Thus, the development and testing of global seismicity models is essential to improving seismic hazard estimates and earthquake-preparedness protocols for effective disaster-risk mitigation. Currently, the availability and quality of geodetic data along plate-boundary regions provides the opportunity to construct global models of plate motion and strain rate, which can be translated into global maps of forecasted seismicity. Moreover, the broad coverage of existing earthquake catalogs facilitates in present-day the calibration and testing of global seismicity models. As a result, modern global seismicity models can integrate two independent factors necessary for physics-based, long-term earthquake forecasting, namely interseismic crustal strain accumulation and sudden lithospheric stress release.
In this dissertation, I present the construction of and testing results for two global ensemble seismicity models, aimed at providing mean rates of shallow (0-70 km) earthquake activity for seismic hazard assessment. These models depend on the Subduction Megathrust Earthquake Rate Forecast (SMERF2), a stationary seismicity approach for subduction zones, based on the conservation of moment principle and the use of regional "geodesy-to-seismicity" parameters, such as corner magnitudes, seismogenic thicknesses and subduction dip angles. Specifically, this interface-earthquake model combines geodetic strain rates with instrumentally-recorded seismicity to compute long-term rates of seismic and geodetic moment. Based on this, I derive analytical solutions for seismic coupling and earthquake activity, which provide this earthquake model with the initial abilities to properly forecast interface seismicity. Then, I integrate SMERF2 interface-seismicity estimates with earthquake computations in non-subduction zones provided by the Seismic Hazard Inferred From Tectonics based on the second iteration of the Global Strain Rate Map seismicity approach to construct the global Tectonic Earthquake Activity Model (TEAM). Thus, TEAM is designed to reduce number, and potentially spatial, earthquake inconsistencies of its predecessor tectonic earthquake model during the 2015-2017 period. Also, I combine this new geodetic-based earthquake approach with a global smoothed-seismicity model to create the World Hybrid Earthquake Estimates based on Likelihood scores (WHEEL) model. This updated hybrid model serves as an alternative earthquake-rate approach to the Global Earthquake Activity Rate model for forecasting long-term rates of shallow seismicity everywhere on Earth.
Global seismicity models provide scientific hypotheses about when and where earthquakes may occur, and how big they might be. Nonetheless, the veracity of these hypotheses can only be either confirmed or rejected after prospective forecast evaluation. Therefore, I finally test the consistency and relative performance of these global seismicity models with independent observations recorded during the 2014-2019 pseudo-prospective evaluation period. As a result, hybrid earthquake models based on both geodesy and seismicity are the most informative seismicity models during the testing time frame, as they obtain higher information scores than their constituent model components. These results support the combination of interseismic strain measurements with earthquake-catalog data for improved seismicity modeling. However, further prospective evaluations are required to more accurately describe the capacities of these global ensemble seismicity models to forecast longer-term earthquake activity.
The plasmasphere is a dynamic region of cold, dense plasma surrounding the Earth. Its shape and size are highly susceptible to variations in solar and geomagnetic conditions. Having an accurate model of plasma density in the plasmasphere is important for GNSS navigation and for predicting hazardous effects of radiation in space on spacecraft. The distribution of cold plasma and its dynamic dependence on solar wind and geomagnetic conditions remain, however, poorly quantified. Existing empirical models of plasma density tend to be oversimplified as they are based on statistical averages over static parameters. Understanding the global dynamics of the plasmasphere using observations from space remains a challenge, as existing density measurements are sparse and limited to locations where satellites can provide in-situ observations. In this dissertation, we demonstrate how such sparse electron density measurements can be used to reconstruct the global electron density distribution in the plasmasphere and capture its dynamic dependence on solar wind and geomagnetic conditions.
First, we develop an automated algorithm to determine the electron density from in-situ measurements of the electric field on the Van Allen Probes spacecraft. In particular, we design a neural network to infer the upper hybrid resonance frequency from the dynamic spectrograms obtained with the Electric and Magnetic Field Instrument Suite and Integrated Science (EMFISIS) instrumentation suite, which is then used to calculate the electron number density. The developed Neural-network-based Upper hybrid Resonance Determination (NURD) algorithm is applied to more than four years of EMFISIS measurements to produce the publicly available electron density data set.
We utilize the obtained electron density data set to develop a new global model of plasma density by employing a neural network-based modeling approach. In addition to the location, the model takes the time history of geomagnetic indices and location as inputs, and produces electron density in the equatorial plane as an output. It is extensively validated using in-situ density measurements from the Van Allen Probes mission, and also by comparing the predicted global evolution of the plasmasphere with the global IMAGE EUV images of He+ distribution. The model successfully reproduces erosion of the plasmasphere on the night side as well as plume formation and evolution, and agrees well with data.
The performance of neural networks strongly depends on the availability of training data, which is limited during intervals of high geomagnetic activity. In order to provide reliable density predictions during such intervals, we can employ physics-based modeling. We develop a new approach for optimally combining the neural network- and physics-based models of the plasmasphere by means of data assimilation. The developed approach utilizes advantages of both neural network- and physics-based modeling and produces reliable global plasma density reconstructions for quiet, disturbed, and extreme geomagnetic conditions.
Finally, we extend the developed machine learning-based tools and apply them to another important problem in the field of space weather, the prediction of the geomagnetic index Kp. The Kp index is one of the most widely used indicators for space weather alerts and serves as input to various models, such as for the thermosphere, the radiation belts and the plasmasphere. It is therefore crucial to predict the Kp index accurately. Previous work in this area has mostly employed artificial neural networks to nowcast and make short-term predictions of Kp, basing their inferences on the recent history of Kp and solar wind measurements at L1. We analyze how the performance of neural networks compares to other machine learning algorithms for nowcasting and forecasting Kp for up to 12 hours ahead. Additionally, we investigate several machine learning and information theory methods for selecting the optimal inputs to a predictive model of Kp. The developed tools for feature selection can also be applied to other problems in space physics in order to reduce the input dimensionality and identify the most important drivers.
Research outlined in this dissertation clearly demonstrates that machine learning tools can be used to develop empirical models from sparse data and also can be used to understand the underlying physical processes. Combining machine learning, physics-based modeling and data assimilation allows us to develop novel methods benefiting from these different approaches.
Near-Earth space represents a significant scientific and technological challenge. Particularly at magnetic low-latitudes, the horizontal magnetic field geometry at the dip equator and its closed field-lines support the existence of a distinct electric current system, abrupt electric field variations and the development of plasma irregularities. Of particular interest are small-scale irregularities associated with equatorial plasma depletions (EPDs). They are responsible for the disruption of trans-ionospheric radio waves used for navigation, communication, and Earth observation. The fast increase of satellite missions makes it imperative to study the near-Earth space, especially the phenomena known to harm space technology or disrupt their signals. EPDs correspond to the large-scale structure (i.e., tens to hundreds of kilometers) of topside F region irregularities commonly known as Spread F. They are observed as depleted-plasma density channels aligned with the ambient magnetic field in the post-sunset low-latitude ionosphere. Although the climatological variability of their occurrence in terms of season, longitude, local time and solar flux is well-known, their day to day variability is not. The sparse observations from ground-based instruments like radars and the few simultaneous measurements of ionospheric parameters by space-based instruments have left gaps in the knowledge of EPDs essential to comprehend their variability.
In this dissertation, I profited from the unique observations of the ESA’s Swarm constellation mission launched in November 2013 to tackle three issues that revealed novel and significant results on the current knowledge of EPDs. I used Swarm’s measurements of the electron density, magnetic, and electric fields to answer, (1.) what is the direction of propagation of the electromagnetic energy associated with EPDs?, (2.) what are the spatial and temporal characteristics of the electric currents (field-aligned and diamagnetic currents) related to EPDs, i.e., seasonal/geographical, and local time dependencies?, and (3.) under what conditions does the balance between magnetic and plasma pressure across EPDs occur?
The results indicate that: (1.) The electromagnetic energy associated with EPDs presents a preference for interhemispheric flows; that is, the related Poynting flux directs from one magnetic hemisphere to the other and varies with longitude and season. (2.) The field-aligned currents at the edges of EPDs are interhemispheric. They generally close in the hemisphere with the highest Pedersen conductance. Such hemispherical preference presents a seasonal/longitudinal dependence. The diamagnetic currents increase or decrease the magnetic pressure inside EPDs. These two effects rely on variations of the plasma temperature inside the EPDs that depend on longitude and local time. (3.) EPDs present lower or higher plasma pressure than the ambient. For low-pressure EPDs the plasma pressure gradients are mostly dominated by variations of the plasma density so that variations of the temperature are negligible. High-pressure EPDs suggest significant temperature variations with magnitudes of approximately twice the ambient. Since their occurrence is more frequent in the vicinity of the South Atlantic magnetic anomaly, such high temperatures are suggested to be due to particle precipitation.
In a broader context, this dissertation shows how dedicated satellite missions with high-resolution capabilities improve the specification of the low-latitude ionospheric electrodynamics and expand knowledge on EPDs which is valuable for current and future communication, navigation, and Earth-observing missions. The contributions of this investigation represent several ’firsts’ in the study of EPDs: (1.) The first observational evidence of interhemispheric electromagnetic energy flux and field-aligned currents. (2.) The first spatial and temporal characterization of EPDs based on their associated field-aligned and diamagnetic currents. (3.) The first evidence of high plasma pressure in regions of depleted plasma density in the ionosphere. These findings provide new insights that promise to advance our current knowledge of not only EPDs but the low-latitude post-sunset ionosphere environment.