This thesis presents my research on the development of a neural network (NN) based global empirical model of the ionospheric F2 region peak electron density using extended geophysically relevant inputs. The main principle behind this approach has been to utilize parameters other than simple geographic co-ordinates, on which the F2 peak electron density is known to depend, and to exploit the technique of NNs, thereby establishing and modeling the non-linear dynamic processes (both in space and time)associated with the F2 region electron density on a global scale. Four different models have been developed in this work. These are the foF2 NN model, M(3000)F2 NN model, short-term forecasting foF2 NN, and a near-real time foF2 NN model. Data used in the training of the NNs were obtained from the worldwide ionosonde stations spanning the period 1964 to 1986 based on availability, which included all periods of calm and disturbed magnetic activity. Common input parameters used in the training of all 4 models are day number (day of the year, DN), Universal Time (UT), a 2 month running mean of the sunspot number (R2), a 2 day running mean of the 3-hour planetary magnetic index ap (A16), solar zenith angle (CHI), geographic latitude (q), magnetic dip angle (I), angle of magnetic declination (D), angle of meridian relative to subsolar point (M). For the short-term and near-real time foF2 models, additional input parameters related to recent past observations of foF2 itself were included in the training of the NNs. The results of the foF2 NN model and M(3000)F2 NN model presented in this work, which compare favourably with the IRI (International Reference Ionosphere) model successfully demonstrate the potential of NNs for spatial and temporal modeling of the ionospheric parameters foF2 and M(3000)F2 globally. The results obtained from the short-term foF2 NN model and nearreal time foF2 NN model reveal that, in addition to the temporal and spatial input variables, short-term forecasting of foF2 is much improved by including past observations of foF2 itself. Results obtained from the near-real time foF2 NN model also reveal that there exists a correlation between measured foF2 values at different locations across the globe. Again, comparisons of the foF2 NN model and M(3000)F2 NN model predictions with that of the IRI model predictions and observed values at some selected high latitude stations, suggest that the NN technique can successfully be employed to model the complex irregularities associated with the high latitude regions. Based on the results obtained in this research and the comparison made with the IRI model (URSI and CCIR coefficients), these results justify consideration of the NN technique for the prediction of global ionospheric parameters. I believe that, after consideration by the IRI community, these models will prove to be valuable to both the high frequency (HF) communication and worldwide ionospheric communities.

Atmospheric planetary waves play a role in introducing variability to the low-latitude ionosphere. To better understand this coupling, this study investigates times when oscillations seen in both atmospheric planetary waves and ionospheric data-sets have similar periodicity. The planetary wave data-set used are temperature observations made by Sounding of the Atmosphere using Broadband Emission Radiometry (SABER). These highlight periods during which 2-Day westward propagating wave-number 3 waves are evident in the mesosphere and lower thermosphere. The ionospheric data-set is Total Electron Content (TEC), which is used to identify periods during which the ionosphere appears to respond to the planetary waves. Data from KP and F10.7 indices are used to determine events that may be of external origin. A 17-year time-span from 2002 to 2018 is used for this analysis so that both times of solar minimum and maximum can be studied. To extract the periods of this collection of data a Morlet Wavelet analysis is used, along with thresholding to indicate events when similar periods are seen in each data-set. Trends are then determined, which can lead to verification of previous assumptions and new discoveries. / Master of Science / The thermosphere and ionosphere are impacted by many sources. The sun and the magnetosphere externally impact this system. Planetary waves, which originate in the lower atmosphere, internally impact this system. This interaction leads to periodic signatures in the ionosphere that reflect periodic signatures seen in the lower atmosphere, the sun and the magnetosphere. This study identifies these times of similar oscillations in the neutral atmosphere, the ionosphere, and the sun, in order to characterize these interactions. Events are cataloged through wavelet analysis and thresholding techniques. Using a time-span of 17 years, trends are identified using histograms and percentages. From these trends, the characteristics of this coupling can be concluded. This study is meant to confirm the theory and provide new insights that will hopefully lead to further investigation through modeling. The goal of this study is to gain a better understanding of the role that planetary waves have on the interaction of the atmosphere and the ionosphere.

In near-Earth space, highly spatio-temporally variant magnetic fields result from solar-terrestrial magnetic interaction. These near-Earth external fields currently represent the largest source of error in efforts to model the magnetic field produced in the Earth’s interior. Starting in 1999, the Decade of Geopotential Field Research (Friis-Christensen et al., 2009) has greatly increased the amount of available low-Earth orbit (LEO) satellite magnetic data. These data have driven many advances in field modelling, yet have highlighted that LEO measurements are particularly susceptible to contamination from external fields. This thesis presents a series of studies attempting to describe the external fields in more detail, in order that they can be more effectively separated from the internal fields in magnetic modelling efforts. A range of analysis methods, different for each study, are applied to satellite and ground-based observatory data. Mandea and Olsen’s (2006) method of estimating the secular variation (SV) of the internal field from satellite data via ‘Virtual Observatories’ (VOs) is applied to synthetic data from the upcoming Swarm constellation satellite mission of the European Space Agency. Beggan (2009) found VOs constructed from CHAMP satellite data to be contaminated with external field signals which appeared to have a significant local time (LT) dependence. I find that utilising the increased coverage of LT sectors offered by the Swarm constellation geometry does not significantly decrease the contamination. Following this surprising result I tested a wide range of methods aimed at reducing the VO contamination from each parameterised external field source region. In anticipation of future studies using real data, I used the results of the tests to provide a more complete description of the external field variations affecting analyses of geographically-fixed magnetic phenomena when using satellite data and spherical harmonic analysis (SHA). Ionospheric electric currents flowing at LEO altitudes are known to violate the assumption of measurements taken in a source-free space, required in SHA-based models of the magnetic field. In order to better describe the electromagnetic environment at LEO altitudes, I use data from the Ørsted and CHAMP satellites to calculate the current density from Amp`ere’s integral. Vector magnetic data from discrete overflights of the two satellites (at different altitudes) are rotated into the along-track frame to define the integral loop and its ‘surface area’, permitting estimation of the predominantly zonal current density flowing in the region between the two orbital paths. I designed selection criteria to extract geometrically-stable overflights spanning the range of LTs twice in the 6 years of mutually available satellite vector data. From these overflights I resolve current densities in the range 0:1 μA=m2, with the distribution of current largely matching the LT progression of the Appleton anomaly. I applied detailed tests to check for biases intrinsic to the method, and present results free of systematic errors. The results are compared with the predictions of the CTIP (Coupled Thermosphere-Ionosphere-Plasmasphere) model of ionospheric composition and temperature, showing a typically good spatiotemporal agreement. I find persistent current intensifications between geomagnetic latitudes of 30 and 50 in the post-midnight, pre-dawn sector, a region which has been previously considered to be relatively free of currents. External fields induce currents in the Earth’s conducting mantle, the magnetic fields of which add to the field measured at and above the Earth’s surface. The morphology of the long-period inducing field is poorly resolved on timescales of months to years, reducing the accuracy of mantle induction studies (a key part of the Swarm mission). I improve the description of its morphology via the method of Empirical Orthogonal Functions (EOFs), which I apply to over a decade of ground-based observatory data. EOFs provide a decomposition of the spatiotemporal structures contained in the magnetic field data, with partitions arising from the data themselves, overcoming the relatively simplistic assumptions made about the inducing field morphology in LT. The results of vector data EOF analyses are presented, but I rely primarily on scalar analyses which are more fitting for this study. I overcome the limitations of the irregular observatory distribution with a novel spatial weighting matrix, combining the output from multiple EOF analyses to greatly improve the data coverage in LT. I find that the seasonal variation of the inducing field is more important than the variation of the symmetric ring current on annual periods, and that dawn-dusk asymmetry should be accounted for to increase the accuracy of mantle conductivity estimates based on data covering the decadal timescales of the solar cycle.

The solar wind, magnetosphere and ionosphere are coupled by magnetohydrodynamic waves, and this gives rise to new and often unexpected behaviours that cannot be produced by a single, isolated part of the system. This thesis examines two broad instances of coupling: field-line resonance (FLR) which couples fast and Alfvén waves, and magnetosphere-ionosphere (MI-) coupling via Alfvén waves. The first part of this thesis investigates field-line resonance for equilibria that vary in two dimensions perpendicular to the background magnetic field. This research confirms that our intuitive understanding of FLR from 1D is a good guide to events in 2D, and places 2D FLR onto a firm mathematical basis by systematic solution of the governing equations. It also reveals the new concept of ‘imprinting’ of spatial forms: spatial variations of the resonant Alfvén wave correlate strongly with the spatial form of the fast wave that drives the resonance. MI-coupling gives rise to ionosphere-magnetosphere (IM-) waves, and we have made a detailed analysis of these waves for a 1D sheet E-region. IM-waves are characterised by two quantities: a speed v_{IM} and an angular frequency ω_{IM} , for which we have obtained analytic expressions. For an ideal magnetosphere, IM-waves are advective and move in the direction of the electric field with speed v_{IM}. The advection speed is a non-linear expression that decreases with height-integrated E-region plasma-density, hence, wavepackets steepen on their trailing edge, rapidly accessing small length-scales through wavebreaking. Inclusion of electron inertial effects in the magnetosphere introduces dispersion to IM-waves. In the strongly inertial limit (wavelength λ << λ_{e} , where λ_{e} is the electron inertial length at the base of the magnetosphere), the group velocity of linear waves goes to zero, and the waves oscillate at ω_{IM} which is an upper limit on the angular frequency of IM-waves for any wavelength. Estimates of v_{IM} show that this speed can be a significant fraction (perhaps half) of the E_{⊥} × B_{0} drift in the E-region, producing speeds of up to several hundred metres per second. The upper limit on angular frequency, ωIM , is estimated to give periods from a few hundredths of a second to several minutes. IM-waves are damped by recombination and background ionisation, giving an e-folding decay time that can vary from tens of seconds to tens of minutes. We have also investigated the dynamics and steady-states that occur when the magnetosphere-ionosphere system is driven by large-scale Alfvénic field-aligned currents. Steady-states are dominated by two approximate solutions: an ‘upper’ solution that is valid in places where the E-region is a near perfect conductor, and a ‘lower’ solution that is valid where E-region depletion makes recombination negligible. These analytic solutions are extremely useful tools and the global steady-state can be constructed by matching these solutions across suitable boundary-layers. Furthermore, the upper solution reveals that E-region density cavities form and widen (with associated broadening of the magnetospheric downward current channel) if the downward current density exceeds the maximum current density that can be supplied by background E-region ionisation. We also supply expressions for the minimum E-region plasma-density and shortest length-scale in the steady-state. IM-waves and steady-states are extremely powerful tools for interpreting MI-dynamics. When an E-region density cavity widens through coupling to an ideal, single-fluid MHD magnetosphere, it does so by forming a discontinuity that steps between the upper and lower steady-states. This discontinuity acts as part of an ideal IM-wave and moves in the direction of the electric field at a speed U = \sqrt{v_{IM} {+} v_{IM} {-}}, which is the geometric mean of v_{IM} evaluated immediately to the left and right of the discontinuity. This widening speed is typically several hundreds of metres per second. If electron inertial effects are included in the magnetosphere, then the discontinuity is smoothed, and a series of undershoots and overshoots develops behind it. These undershoots and overshoots evolve as inertial IM-waves. Initially they are weakly inertial, with a wavelength of about λ_{e}, however, strong gradients of ω_{IM} cause IM-waves to phase-mix, making their wavelength inversely proportional to time. Therefore, the waves rapidly become strongly inertial and oscillate at ω_{IM}. The inertial IM-waves drive upgoing Alfvén waves in the magnetosphere, which populate a region over the downward current channel, close to its edge. In this manner, the E-region depletion mechanism, that we have detailed, creates small-scale Alfvén waves in large-scale current systems, with properties determined by MI-coupling.

The nighttime reflection height of low-frequency (LF) radio waves at oblique incidence is measured at Collm Observatory using 1.8 kHz sideband phase comparisons between the sky wave and the ground wave of a commercial 177 kHz LF transmitter. The measurements have been carried out continuously since September 1982, now allowing the analysis of trends and regular variations of the reflection height. In the time series is found a) a long-term negative trend and b) a solar cycle dependence, both confirming earlier measurements and theoretical estimations. Moreover, a significant oscillation of quasi-biennial period is visible in LF reflection heights, indicating a reaction of the midlatitude mesosphere/lower thermosphere region on the equatorial quasi biennial oscillation. / Am Observatorium Collm der Universität Leipzig werden die nächtlichen Reflexionshöhen von Langwellen auf 177 kHz unter Verwendung von Seitenbandinformationen in einem kleinen Bereich um 1.8 kHz gemessen. Die Messungen werden seit September 1982 durchgeführt und erlauben nunmehr Analysen von Langzeittrends und regulären Variationen der unteren Ionosphäre. Bei der Untersuchung der Zeitreihen fallen die folgenden Zusammenhänge auf: a)es existiert ein negativer Trend, welcher mit der Abkühlung der Stratosphäre in Zusammenhang steht, b) die Reflexionshöhe weist eine Modulation mit dem 11-Jährigen Sonnenfleckenzyklus auf und c) es ist eine deutliche quasi-zweijährige Schwingung sichtbar, die auf eine Kopplung der Mesosphäre und unteren Thermosphäre mit der äquatorialen Stratosphäre hinweist.

Während der Sonnenfinsternis am 11. 8. 1999 kam es zu einer kurzzeitigen starken Abnahme der Ionisation im Höhenbereich der D-Region (60-90 km Höhe) am Unterrand der Ionosphäre. Mit bodengebundenen funktechnischen Beobachtungen der Ausbreitung elektromagnetischer Wellen über die D-Region konnte dieses Ereignis in charakteristischer Weise als ein simulierter Tag-Nacht-Übergang mit allen zu erwartenden Konsequenzen beobachtet werden. Auf der Basis von Funkwellen-Ausbreitungsmessungen werden in Collm (51.3°N, 13°E) Windmessungen im Höhenbereich der Mesopause und unteren Thermosphäre (80-105 km) durchgeführt, die in der Regel - bedingt durch die Besonderheiten der ionosphärischen Wellenausbreitung des verwendeten Frequenzbereiches - nur in den Nachtstunden möglich sind. Während der Hauptphase des Finsterniseffektes wurden Messungen auch am Tag möglich. Die Ergebnisse fügen sich gut in das aus den mittleren monatlichen Nachtwerten für die Tagstunden extrapolierte Niveau ein. Die Reaktion der Ionosphäre auf die Sonnenfinsternis erfolgte mit geringer Verzögerung: der maximale Effekt war 5-10 Minuten nach der größten Abdeckung der Sonnenscheibe zu verzeichnen. / During the solar eclipse on 11 August 1999 a short-term decrease of ionisation in the D-region ( 60- 90 km altitude) at the lower boundary of the ionosphere appeared. Using ground-based measurements of radio-wave propagation through the D-region the event could be monitored as a simulated day-night-transition with the expected consequences for the ionosphere. At Collm (51.3°N, 13°E), mesopause region (80-105 km) windmeasurements are carried out based on radio wave propagation, which are not possible in summer during daylight hours due to radio-wave absorption. During the eclipse measurements became possible for a short period. Tue results fit well to mean monthly daytime data extrapolated from nighttime measurements. The reaction of the ionosphere on the solar eclipse was fast; the maximum effect was registered only 5-10 minutes after the maximum eclipse of the solar disk.

Using the spaced receiver method in the low-frequency (LF) range, lower E-region ionospheric drifts are measured at Collm Observatory, Germany since several decades. These drifts are interpreted as upper mesospheric winds at the reflection height of the used amplitude modulated LF radio waves, the latter being measured since 1983 using travel time differences between the ground wave and the ionospherically reflected sky wave within a small sideband range near 1.8 kHz above and below the carrier frequency. One regular feature of midlatitude upper mesosphere winds is the quasi twoday wave (QTDW), known as a wavenumber 3 or 4 wave in the middle atmosphere, usually occurring as one or more bursts during the summer season at midlatitudes. The OTDW bursts, as measured in LF winds, shows substantial decadal and interannual variability. Comparison with the background winds show that the onset of QDTW bursts is found near maximum values of the vertical wind shear, and maximum QTDW amplitudes are measured, on average, about one week after the maximum wind shear. This supports the theory that the QTDW is forced by instability of the summer mesospheric wind jet. / Am Observatorium Collm werden seit mehreren Jahrzehnten Langwellenwindmessungen in der unteren ionosphärischen E-Schicht durchgeführt. Die zugehörige Reflexionshöhe wird, auf der Basis von Laufzeitdifferenzmessungen zwischen der Raum- und Bodenwelle, seit 1983 ebenfalls registriert. Eines der regelmäßig beobachteten Phänomene ist die quasi 2-Tage-Welle, die als eine planetare Welle der Wellenzahl 3 oder 4 bekannt ist. Diese Welle erscheint in mittleren Breiten in einem oder mehreren Schüben im Sommer. Nach den Messungen am Collm besitzt die Welle eine deutliche Variabilität von Jahr zu Jahr. Vergleiche mit dem zonalen Grundwind zeigen, dass das Auftreten von Maxima der 2-Tage-Welle in vielen Fällen mit erhöhter vertikaler Windscherung in Verbindung steht, so dass im langzeitlichen Mittel maximale Wellenamplituden einige Tage nach dem Auftreten maximaler Windscherung zu finden sind. Dies unterstützt die These, dass die quasi 2-Tage-Welle durch barokline Instabilität des sommerlichen Mesosphärenjets angeregt wird.

Globale Karten des totalen Elektronengehaltes (TEC) der Ionosphäre werden nach Signalen planetarer Wellenaktivität aus der Stratosphäre im Bereich der mittleren Breiten (ca. 52.5° N) untersucht, um eine Abschätzung über die vertikale Kopplung durch planetare Wellen (PW) zu erhalten. Die Variabilität der Ionosphäre wird operationell durch das DLR Neustrelitz erfasst. Seit 2002 werden zu diesem Zwecke hemisphärische TEC Karten erstellt, die eine Analyse PW typischer Oszillationen in der Ionosphäre ermöglichen. Die verwendete Methode zur Analyse separiert Wellen nach ihrer zonalen Wellenzahl, Periode und Ausbreitungsrichtung. In einer vorherigen Fallstudie vom Herbst 2004 wurde u.a. die quasi 6-Tage Welle (m2w) im mittleren Spektrum für das Geopotential in 1hPa (Stratosphäre) als auch den ionosphärischen TEC beobachtet. Die aktuellen Resultate geben Hinweise für ein gleichzeitiges Auftreten dieserWelle mit einer quasi 6-Tage Oszillation in der Mesopausenregion. Jedoch im Vergleich zur Stratosphäre scheinen die Signaturen verschoben und etwas modifiziert. / The response of stratospheric planetary wave (PW) activity over the higher middle latitudes (ca. 52.5° N) in global gridded ionospheric data of the total electron content (TEC) are investigated to estimate the vertical coupling by PW. The monitoring of ionospheric variability is regularly operated by DLR Neustrelitz since 2002 producing TEC maps covering the northern hemisphere. This data base is considered for comparing simultaneous observations of wave activity in both stratosphere and ionosphere. The analysis technique of planetary wave type oscillations (PWTO) is carried out by separating waves into their zonal wavenumber, period and travelling direction. A previous case study of autumn 2004 has shown that among other things the quasi 6-day wave (m2w) is visible in the mean spectrum of stratospheric geopotential height at 1 hPa pressure level and of ionospheric TEC data. The actual results give hints for a simultaneous occurrence of this wave type with a quasi 6-day oscillation in the mesopause region. But in comparison to the stratosphere, the wave signatures seem to be somewhat schifted and modified.

O posicionamento de pontos, com o sistema GPS, tornou-se uma ferramenta importante, aplicável nas mais diferentes áreas do conhecimento. No entanto, em algumas situações, a exigência de elevadas precisões trouxe a inconveniência de um custo elevado na aquisição de receptores de dupla freqüência. Mesmo com os avanços tecnológicos, a ionosfera é uma das fontes de erro que mais afetam o posicionamento de pontos. Para os usuários que possuem equipamentos de dupla freqüência, este erro é modelado com grande eficiência através do processamento de dados com solução iono free. No Brasil, a maioria dos usuários possui equipamentos que captam informações apenas do código C/A e/ou código C/A e portadora L1. Neste caso faz-se uso de alguns modelos, como, por exemplo, o de Klobuchar, com redução do erro nos posicionamentos absolutos de, aproximadamente, 50%. Nos posicionamentos relativos, com bases superiores a 20 Km, a utilização deste modelo não é a mais indicada. Neste contexto, este trabalho consiste no desenvolvimento de um método que possibilite modelar o atraso ionosférico através de um polinômio do segundo grau, baseado no ajustamento seqüencial de observações. As informações necessárias para esta modelagem são advindas das estações GPS da Rede Brasileira de Monitoramento Contínuo. Isto possibilita, em posicionamentos absolutos, determinar o atraso ionosférico de uma forma mais eficiente que o obtido pelo modelo de Klobuchar. Em posicionamentos relativos, para os usuários de equipamentos de simples freqüência, tal modelagem permite a geração de um código com características semelhantes ao código P2. Assim, com os dados gerados por receptores de uma freqüência, é possível processar vetores de bases longas por meio da solução iono free code. Os resultados obtidos indicam que tal metodologia pode ser uma alternativa eficiente para minimizar o efeito ionosférico no posicionamento de pontos com o sistema GPS. Horizontalmente, através dos métodos de posicionamento Single Point e relativo, respectivamente, o modelo ionosférico proposto proporcionou uma melhoria de 39% e 26% se comparado com o modelo de Klobuchar. / Point positioning, with GPS, became an important tool applicable to the most different areas of the knowledge. However, in some situations, the requirement of high precisions brought the inconvenience of a high cost in the acquisition of dual receivers frequency. Despite of the technological advances, the ionosphere is one of the error sources that affect most point positioning. For users who have dual equipment frequency, this error is modeled with great efficiency through data processing with the ionosphere free solution. In Brazil the majority of the users has equipments that process C/A code information only and/or C/A code and L1 carrier. In this case it is necessary to use some models, for example the Klobuchar model, with error reduction of approximately 50% in absolute point positioning. In the relative positioning, with baselines longer than 20 Km, the use of this model is not indicated. Thus, this work consists in developing a method that makes possible to model the ionospheric delay using a second degree polynomial, based on sequential adjustment of observations. The necessary information for this modeling is obtained from GPS stations that compose the RBMC (Brazilian Network for Continuous Monitoring of GPS). Thus it is possible to determine the ionospheric delay in a more efficient way in absolute positioning than when using the Klobuchar model. In relative positioning, for single frequency users, such modeling allows the generation of a code, with similar characteristics to the P2 code. Hence, with data generated by single frequencies, it is possible to process long base line vectors, using the iono free code solution. The results obtained indicate that such methodology can be an efficient alternative to minimize the ionospheric effect in the GPS point positioning. Horizontally, through the methods of positioning Single Point and relative, respectively, the ionospheric model considered provided an improvement of 39% and 26% if compared with the Klobuchar model.

This thesis covers two main areas which are related to the mapping and examination of the ionosphere. The first examines the performance and specific nuances of various state-of-the-art interpolation methods with specific application to mapping the ionosphere. This work forms the most widely scoped examination of interpolation technique for ionospheric imaging to date, and includes the introduction of normalised convolution techniques to geophysical data. In this study, adaptive-normalised convolution was found to perform well in ionospheric electron content mapping, and the popular technique, kriging was found to have problems which limit its usefulness. The second, is the development and examination of automatic data-driven motion estimation methods for use on ionospheric electron content data. Particular emphasis is given to storm events, during which characteristic shapes appear and move across the North Pole. This is a particular challenge, as images covering this region tend to have a very-low resolution. Several motion estimation methods are developed and applied to such data, including methods based on optical flow, correlation and boundarycorrespondence. Correlation and relaxation labelling based methods are found to perform reasonably, and boundary based methods based on shape-context matching are found to perform well, when coupled with a regularisation stage. Overall, the techniques examined and developed here will help advance the process of examining the features and morphology of the ionosphere, both during storms and quiet times.