This thesis was carried out in two parts. The stellar spectral data was used from the Gaia-ESO survey. The data used was fromthe public archive as well as data received from Dr. Recio-Blanco at Observatoire Cote D'Azure. 1) I performed anomaly detection using unsupervised random forests, by applying the concept of weirdness scores to identify outliers. 2) Using spectral data along with physical parameters of objects in the galactic bulge of the Gaia-ESO survey, I built a variational autoencoder neural network to reconstruct stellar spectra and explore latent features learning physical parameters by themselves.

SN2020qlb (ZTF20abobpcb) is an extensively sampled hydrogen-poor superluminous supernova (SLSN-I) that is among the most luminous (max. Mg = -22:25+/-0:01 magnitudes) and long-rising (77.1 days from explosion to maximum) in a category of the brightest and longest rising SNe currently known in the Universe. SN2020qlb exhibits clear light-curve undulations, a phenomenon seen in other SLSNe but whose physical origin is still a mystery. This Master of Science Thesis discusses both the potential power source of these immense explosions as well as the power mechanisms behind the observed light-curve undulations. A particularly large set of photometric data in both the visible and ultraviolet ranges and covering the first 410 days after explosion, as well as 10 spectra are available for analysis. The explosion date is constrained to +/-0.28 days; the phase and magnitude of the peak luminosity are determined; light-curves for each telescope/filter combination are constructed; the g-r color evolution is plotted; the photospheric temperature and radius evolutions are estimated; the bolometric light-curve is constructed and compared with known power source models; and the rest-frame spectral evolution is plotted for analysis. SN2020qlb is found to meet all the known criteria to be a SLSN-I. A radioactive power source model based on the decay of 56Ni is rejected due to unphysical parameter results. A source model based on the dipole spindown energy deposition of a magnetar fits well the bulk portion of the bolometric light-curve with physically possible parameter values. Two full periods of about 32+/-6 day undulations are found on top of the bulk light-curve after subtracting the smooth model values. Evidence for them is also seen in each filter light-curve. Hypothetical power sources for these striking characteristics from the literature are analyzed and discussed. In summary, a magnetar source for the bulk of the light curve is favored. An external source for the undulations is favored, e.g. interaction with variations in the progenitor star’s circumstellar material, with a caveat that an explanation involving the break-down of model assumptions cannot be ruled out. Together they are favored to explain the entire light-curve of SN2020qlb.

Failed supernovae are the implosive final fates of massive stars, where ablack hole is formed. During the collapse, the proto-neutron star emits a huge number of neutrinos, and when the black hole is finally formed, it engulfs theneutrino-emitting material and the signal is cut off. Inspired by the recent work of Nagakura & Hotokezaka (2020), this thesis improves on some parts of theirs imulation work and further explores the neutrino signal from failed supernovae, using a supercomputer to perform Monte Carlo simulations. In particular, we realized the neutrino flux’ time evolution around black hole formation hasn’t previously been studied well, and so it is investigated here, as well as the plausibility of measuring the black hole mass through the shape of the decay. A new component of the signal is presented, an echo of neutrinos emitted before black hole formation that, due to scattering in supersonic material around the black hole, arrive with a time delay of up to 15 ms, and with a significantly higher average energy, for heavy lepton neutrinos around 50 MeV.

Modelling stellar structures and comparing them with observationsis a very important step when it comes to verifying our theories aboutstellar evolution. Three-dimensional reconstruction is therefore impor-tant and in the case of certain stellar types it makes for a large portionof the ongoing research.For this project, three dierent objects and their three-dimensionalmodels were selected for 3D-printing. The systems in question areEta Carinae, 1 Gruis and HD 101584 and the reason behind thischoice is the fact that each object showcases a dierent phase and/orprocess of stellar evolution approaching or during the planetary nebula(PN) stage. On top of that, these objects have been observed using atechnique that allows us to deduce their 3D structure.The three-dimensional models and prints allow us to nd features(such as axial symmetries) that give us more information about themovements inside the system and their consequences on how the stellarstructure has and will evolve.

For this project, a historical statistical analysis of the Amplitude of Mass Fluctuations ($\sigma_8$) and Hubble's Constant ($H_0$) parameters in the Standard Cosmological Model was carried out to determine whether or not the given error bars truly represent the dispersion of values. It was found through analysis of the Chi-Squared ($\chi^2$) values of the data that for $\sigma_8$ (60 data points and $\chi^2$ between 183.167 and 189.037) that the associated probability Q is extremely low, with $Q = 1.5597*10^{-15}$ for the weighted average and $Q = 1.2107*10^{-14}$ for the best fit of the data. This was also the case for the $\chi^2$ values (163 data points and $\chi^2$ between 484.3977 and 575.655) of $H_0$, where $Q = 4.2176*10^{-34}$ for the linear fit of the data and $Q = 1.0342*10^{-47}$ for the weighted average of the data. Through further analysis, it is shown in question, a linear fit is a better estimate of the data than the weighted average. The general conclusion is that the statistical error bars have been underestimated (in around 20\% of the measurements), or the systematic errors were not properly taken into account.

This project aims to determine the solar Mg abundance with two infrared spectral lines: MgI λ10811 and λ10965. Downloaded line data from VALD are updated with modern values for the oscillator strengths and van der Waals damping parameters, the latter obtained through ABO theory. Utilizing SME, the Mg abundances of synthetic spectra are fitted with respect to a solar atlas. The derived abundance for varied turbulence configurations is found to be between logAMg+12 = 7.40–7.52, which is slightly lower than meteoric and 3D-modeled values. Suggested improvements would be to consider the effects of NLTE and line blending.

The current cosmological paradigm of an accelerating cosmic expansion issupported by observations of Type Ia supernovae. However, the light emittedby these and other cosmological sources is not only redshifted by cosmicexpansion but will also interact with matter along the light path. Especiallyintergalactic dust can lead to additional reddening and dimming of distantsources due to light scattering or absorption. This yields systematiccontaminations to cosmological measurements. This project builds afoundation and some of the tools that will be used in a master’s thesis withthe aim of analysing the spatial distribution and the properties of this cosmicdust. While previous studies assumed cosmic dust to be homogeneouslydistributed, it is expected to follow the spatial distribution of galaxies fromwhich it was expelled. This project also starts to recreate previous models ofhomogeneous dust models and measurements which will be used the futuremaster’s thesis. An analysis of the methods and tools used, along with some ofthe dust properties, is also made in this project.

The complex interactions between the Sun and Earth are referred to as Space Weather. Key parameters include magnetic indices which quantitatively describe geomagnetic activity by determining a baseline that removes the background magnetic field and allows quantification of the remaining activity during geomagnetic events. However, most used indices have a low temporal resolution and rely on a sparse and frozen network of ground magnetic observatories. This thesis introduces a novel way of determining the baseline for future high temporal and spatial resolution magnetic indices. Firstly, the main phenomena and effects of Space Weather are outlined, followed by a review of currently used magnetic indices and their derivation. The computation of a novel baseline introduced in this work relies on basic statistical methods which are applied on magnetic data from a dense and flexible network of ground observatories for the period 1991-2016. The focus is on the investigation of geomagnetic quiet periods for which average annual activity at each observatory is determined. A global latitudinal normalisation function with dependency on solar activity for quiet periods is found. The analysis of the newly derived baseline shows that it provides the temporal, spatial and amplitudinal resolution needed to characterise geomagnetic disturbances adequately. The residual signal has the capability of being used as the basis for further quiet period studies. A first attempt of new indices based on the introduced derivation shows a good agreement with already existing high temporal and spatial resolution magnetic indices. Future indices derived with this baseline lay a favourable fundament for the application of articial intelligence methods.

20 supernovae that spectroscopically match the peculiar, superluminous type Ia supernova 2003fg are studied in this project. SN2003fg is thought to have erupted at a super-Chandrasekhar mass, thus breaching the theoretical mass limit for a white dwarf. By analysing the lightcurves of these 20 supernovae, this work aims to understand what the progenitor binary systems from which the supernovae erupt looked like. A lightcurve fitting using the software snpy is performed for each supernova. Using the produced models, time of maximum luminosity, stretch and maximum magnitudes in the g-, r- and i-bands are found. It is found that subluminous supernovae might be a sign of circumstellar material surrounding the progenitor star, though some of the supernovae were superluminous and some adhered to Phillip's relationship. Substructures were found in the lightcurves, as the sampled supernovae showed clearly different behaviours in each of the three bands.

Localized structures of the magnetic field strength depression are often observed in the interplanetarymedium, and they are called ‘magnetic holes’ after the original work of Turner et al. 1977. A numberof observations of similar features have been reported, while the mechanisms of their origin have notfully understood yet. The scale size of their structures varies from several to a few thousand of the protongyro radii, and their characteristic orientations of the magnetic field also vary, and therefore differenttypes of the magnetic holes have been suggested. To date, the magnetic holes are classified into Mirrormode and magnetic decreases (Tsurutani et al. 2011). Despite a large number of papers that report theobservational characteristics of the magnetic holes, many identify the feature using only the magneticfield data. This is due to the scale size of the structure at a large speed of the solar wind medium, thespatial resolution of the plasma instruments is often insufficient while the magnetic field instrument canusually obtain the data in high enough time resolution.The Cassini spacecraft orbited Saturn for almost 17 years and obtained a large amount of data in/near theKronian magnetosphere, where the series of the magnetic depletions have been also observed (Smith et al.1980). The Langmuir Probe (LP) onboard Cassini measures the spacecraft potential and, in turn, measuresthe electron density in in-situ in the outer magnetosphere and solar wind region. This measurement hasbeen done using the LP sweep mode which samples the current-voltage curve of the probe every 10 minin the outer magnetosphere. The LP has also been operated in the continuous mode that measures theprobe current at a fixed bias potential every 16 s allowing to calculate the electron density in a smallerscale that is required for the studies magnetic holes. However, there is no general calibration so far inorder to conduct a statistical study in the outer magnetosphere region. The goal of this project is toinvestigate the possibility to use the LP data for the magnetic hole study, calibrate the LP continuousmode to derive the plasma density near the magnetospheric of Saturn, and investigate the scale size of theplasma density structure in the magnetic holes, i. e. plasma β, the field strength and density.The calibration of the continuous data was done by finding a relation between the current at 11 V, whichis a typical bias voltage of continuous mode, and the spacecraft potential obtained by the LP the sweepmode data. Is is expected that the current at 11V is linearly proportional to the floating potential andtherefore can be used to derive the electron density with the potential and density relationship presentedby Morooka et al. 2009. I found that the spacecraft attitude against the sun has a strong effect on therelation, and derived 11V current-floating potential relationship depending on the different spacecraftattitude.Using the LP continuous data calibration above, I investigated the electron density characteristics aroundthe magnetic hold structure, and confirmed that they are generally in anticorrelation relationship. I estim-ated also the plasma β assuming a constant temperature of 100 eV and investigated their characteristicsfor the different types of magnetic holes (linear and rotational holes) both in the magnetosheath and theholes in the solar wind for the year 2011. For the Cassini dataset during 2011, various different shapeand sizes of magnetic hole events have been found. Most (80%) of the MHs appeared within a groupedstructure, while the rest (20%) are found as isolated type holes in the magnetosheath. Among the isolatedMHs, about 30% had "Gaussian shape" and about 40% had a substructure. The scale size for the electrondensity for the isolated holes were on average 50 s in the solar wind, and 75 s (the rotational holes) and120 s (the linear holes) in the magnetosheath. Therefore, I confirmed that the LP can obtain enough datapoints to resolve the magnetic holes structure in the magnetosheath. The Cassini LP data resolution isalso capable to resolve some of the magnetic hole structure in the solar wind.In summary, I confirmed that the Cassini LP continuous data calibrated in this study is capable toinvestigate the different types of magnetic hole structures. Using this calibrated electron data statisticallyfor the large number of Cassini orbit would helpful to further identify the MHs structures in the solar wind and the magnetosheath that can be a key to understand the generation mechanisms of the magneticholes. / Lokaliserade strukturer med låg magnetfältstyrkan ses ofta i interplanetära mediet och de kallas ’mag-netiska hål’ (MH) (Turner et al. 1977). Trots et antal observationer av sådana strukturer har observeratsär deras generationsmekanism ännu förstådd. Storleken av strukturerna varierar från ett fåtal till någratusen protongyroradier och även deras kännetecknande inriktningar i magnetfältet varierar. På grund avdetta har olika typer av MH förslagits. Idag klassificerar man MH som ’mirror mode’ och magnetiskaminskningar (Tsurutani et al. 2011). Många studier har undersökt de magnetiska hålens egenskaper,men tyvärr oftast baserats endast på magnetfältsdata. Detta kan bero på strukturernas storlek vid en storsolvindshastighet, där plasmainstrumenten oftast inte har tillräckligt hög tidsupplösning för mätningar,medan magnetfältsinstrumenten kan oftast tillhandahålla data i hög tidsupplösning.Cassini-rymdfarkosten kretsade runt Saturnus i nästan 17 år och erhöll stora mängder data i och näraSaturnus magnetosfär. Langmuir-sonden (LP) ombord Cassini mäter rymdfarkostens potential ochdärmed mäter den elektrontätheten i rymden. Instrumentet fungerar som en slags väderstation för rym-dplasma och möjliggör mätningen av fundamentala plasmaparametrar såsom elektrontäthet, jontäthet,elektrontemperatur och jonmassa i en tät plasmaområdet av nära Saturnus. I den yttre magnetosfären därden plasmatätheten är låg, kan LP mäta rymdfarkosts potential och plasmatätheten. Mätningen, så kallade’sweep mode’ kan skaffades var 10:e minuter. LP:en mäter också i ’kontinuerlig mode’ som möjligenkan mäta plasmatätheten i mer frekventa men den behöver allmän kalibrering. I detta projekt undersökerjag möjligheten att använda LP kontinuerlig data för att studera MH, skapa kalibraring funktion för’kontinuerlig mode’ för att uppskatta plasmatätheten i Saturnus magnetosfär, och även att undersökastorleken och karaktär av plasmatäthetenstrukturen i MH.Jag undersökte först relationen mellan LP ström vid 11V och rymdfarkostens potential i sweep mode data.De härledda funktionerna användes vidare för att uppskatta densiteten med användning av relationenmellan rymdfarkostens potential och elektrontätheten (Morooka et al. 2009). Jag upptäckte också attden kontinuerlig mode funktionen är olika beroende på LP sensors läge i förhållande till solen ochrymdfarkosten. Hur Cassini är inriktad har en stor effekt på relationen och därför beskriva jag fyra olikarelationer för olika inriktningsregioner. Med den kontinuerlig mode funktionen jag härlett, undersöktejag struktur av magnetiska hålen som har listats av Tomas Karlsson på KTH. År 2011 innehåller MH medmycket olika former och storlekar. Den mest (80%) MH identifierades som grupp och resten (20%) varsom isolerade MH i magnetosheath. Av dessa isolerande hål har ca. 30% en Gauss-form och nästan 40%av MH verkar ha en understruktur. Genom att jämföra magfältdatan med elektrontätheten bekräftadejag den allmänna antikorrelationen mellan magnetfältstyrkan och elektrontätheten i MH-strukturerna.Dessutom hittar jag en ökning av elektron β som beräknas med en temperatur av 100 eV för linjära ochroterade MH i den magnetosheath samt MH i solvinden under 2011. Storleken av de isolerade magnetiskahålen är i genomsnitt 50 s i solvinden, 75 s (roterade magnetiska hål) och 120 s (linjära magnetiska hålen)i magnetosheath:en. Därför kan Cassini LP ha tillräcklig många datapoäng för att upplösa struktur avMH i magnetosheath. I solvinden kan LP upplösa en del av relativt stora MH.Sammanfattningsvis kan LP:s kontinuerlig kalibreringen från detta projekt användas för att analyserade olika strukturerna och storlekar av MH. Med denna kalibrerade plasmatäthet data är det möjligt attidentifiera olika MH karaktär i statistiskt för det stora antalet Cassini data. Det skulle vara en stor hjälpför att förstå genereringsmekanismerna av de magnetiska hålen.