Browsing by study line "Astrophysical Sciences"

The aim of this thesis is to explore applications of machine learning to the study of asteroid spectra, and as such, its research question can be summarized as: How can asteroid spectra be analyzed using machine learning? The question is explored through evaluation of the obtained solutions to two tasks: the optimal locations of spectrophotometric filters for asteroid classification success and the formation of an asteroid taxonomy through unsupervised clustering. First, background theory for asteroids and particularly spectroscopy of asteroids is presented. Next, the theory of machine learning is briefly discussed, including a focus on the method utilized to solve the first task: neural networks. The first task is executed by developing an optimization algorithm that has access to a neural network that can determine the classification success rate of data samples that would be obtained using spectrophotometric filters at specific locations within the possible wavelength range. The second task, on the other hand, is evaluated through determining the optimal number of clusters for the given dataset and then developing taxonomies with the clustering algorithm k-means. The obtained results for the first task involving the optimal locations of filters for spectrophotometry seem reliable, and correlate relatively well with well-known mineralogical features on asteroid surfaces. The taxonomic systems developed by the unsupervised clustering also succeeded rather well, as many of the formed clusters seem to be meaningful and follow the trends in other asteroid taxonomies. Therefore, it seems that based on the two investigated tasks, machine learning can be applied well to asteroid spectroscopy. For future studies, larger datasets would be required for improving the overall reliability of the results.

The solar corona constantly emits a flow of charged particles, called the solar wind, into interplanetary space. This flow is diverted around the Earth by the magnetic pressure of the Earth’s own geomagnetic field, shielding the Earth from the effect of this particle radiation. On occasion the Sun ejects a large amount of plasma outwards from the corona in an event called a Coronal Mass Ejection (CME). Such events can drive discontinuities in the solar wind plasma, called interplanetary shocks. Shocks can affect the Earth’s magnetosphere, compressing it inwards and generating electromagnetic waves inside it. In this thesis we will cover a study of the ultra-low frequency (ULF) wave response in the magnetosphere to CME-driven shocks. Geomagnetic pulsations are ultra-low frequency plasma waves in the magnetosphere, observable from ground-based magnetometers. The compression of the magnetosphere by interplanetary shocks generates geomagnetic pulsations in the Pc4 and Pc5 frequency ranges (2 - 22 mHz). These waves play an important role in magnetospheric dynamics and the acceleration and depletion of high energy electrons in the radiation belts. We consider 39 interplanetary shock events driven by CMEs, and analyse ground-based magnetometer data from stations located near local noon at the time of the shock arrival. Solar wind measurements are used to categorise interplanetary shocks based on their Mach number and the dynamic pressure differential as main indicators of shock strength. The importance of these parameters in determining the strength of the wave response in the geomagnetic field is then studied using wavelet analysis and superposed epoch analysis. Stronger shocks are found to result in larger increases in wave activity, especially in the Pc4 range. Ground stations at higher latitudes observe higher wavepower, but there is an interesting anomaly in the Pc4 range at stations magnetically connected to regions near the plasmapause, which show an enhanced wavepower response. We quantify the decay time of the wave activity and find that it is around 20 hours for Pc5 waves and 7 hours for Pc4 waves.

Magnetosheath jets are a class of structures in the Earth's magnetosheath usually defined by an enhancement of the dynamic pressure of the plasma. Magnetosheath jets have been observed by several different spacecraft over the past few decades, but their origin and formation mechanism have remained unclear. The aim of this thesis is to use data from a global simulation to investigate the origin of magnetosheath jets. We defined two different kinds of structures, magnetosheath jets and foreshock compressive structures (FCS), and collected a database of individual jets and FCSs from 4 Vlasiator global hybrid-Vlasov simulation runs, all of which simulate only the ecliptic plane. We then conducted a statistical analysis of the properties of jets and FCSs, and their occurrence rates as a function of the definition of the FCS criterion. Jets were separated into two categories: jets that form in contact with FCSs (FCS-jets), and those that do not (non-FCS-jets). We found that up to 75% of magnetosheath jets form in association with an FCS impacting the Earth's bow shock. We also found that FCS-jets penetrate deeper into the magnetosheath than non-FCS-jets. Finally, we found no conclusive explanation for the formation of non-FCS-jets. The properties of both jets and FCSs agree qualitatively and to some extent quantitatively with spacecraft observations and other simulations in the literature. The formation of jets from FCSs impacting the bow shock is similar to the proposed theory that jets are linked to Short Large-Amplitude Magnetic Structures (SLAMS). In the future, we will study magnetosheath jets and FCSs in polar plane simulation runs as well, and ultimately in full 3D simulation runs. If made possible by new simulations, the effects of electron kinetic effects on jets and FCSs will also be studied. Comparison studies with spacecraft observations of jet formation from FCSs will also be conducted, if and when such observations are found and become available.

Mars is a rocky planet in the Solar System, fairly similar to the Earth. It is known for its red color and the theory that there has been liquid water and possibly life in Mars at some point in its history. Mars has been a target of study since space exploration began, with the first fly-by mission occuring in 1965. In the past, Martian climate could have supported a functioning hydrological system and Mars even could have had an ocean. This system would have been similar to the one happening in Antarctica due to cold weather. The Martian atmosphere is mostly composed of carbon dioxide, much thinner and colder compared to the Earth's atmosphere. Methane is a main point of interest due to it being a possible biosignature, a sign of life. Another interesting feature about the atmosphere are the dust devils that contibute to the climate by lifting dust into the air. The Martian soil is composed of rocks and dust containing toxic perchlorates. To colonize Mars, several requirements need to be met: the transportation and funding, designing the settlement efficiently with everything needed and determining how many people are necessary. The risk assessment to both people and equipment needs to be made and taken into account.

Mercury is the smallest planet in the Solar System. The planet has a tenuous atmosphere, which means that it can be modeled as an atmosphereless object. Such Solar System bodies are covered in loose material called regolith, which affects how the object scatters light from the Sun. Photometry is a type of measurement that records the intensity of scattered light as a function of the viewing geometry, which is defined using angles from the surface towards the Sun and towards the observing instrument. Mercury was studied in 2011–2015 by the MESSENGER (MErcury Surface, Space ENvironment, GEochemistry and Ranging) mission of the United States’ National Aeronautics and Space Admin- istration, NASA. The present thesis uses spectrophotometric data, i.e., brightness as a function of wavelength, from the spacecraft’s Mercury Dual Imaging System (MDIS) instrument. Two theo- retical models, the Lommel–Seeliger (LS) and particulate medium (PM) models, are fitted to the observed reflectance using the least-squares method. The PM model is the more complicated of the two, and it includes a shadowing correction that de- pends on three model parameters. The parameters describe the properties of a particulate medium, i.e., regolith. The most important of the parameters is the packing density of the regolith, which is defined as the ratio of the volume of the particles to the total volume. The other two parameters describe the medium’s surface roughness in horizontal and vertical directions. The PM model is fitted to the observed reflectance for various different combinations of the model parameters. Ini- tially, only discrete and predetermined parameter values are used, but the parameter values are extended to arbitrary values using interpolation. Trilinear interpolation is utilized using several methods, followed by Markov chain Monte Carlo (MCMC) sampling for the final results. Most of the methods agree with one another, and fall within the uncertainties of the best solution, which allows to form an argumented conclusion about the best parameter values of the PM model. The best parameter values correspond to a densely-packed regolith with horizontally smooth surface and large height variations. The results of the present study can be used in the BepiColombo mission to Mercury, which is planned to begin its science operations in early 2026.

Active galactic nuclei (AGN) are one of the most powerful sources of the luminous Universe. Radio-loud AGN exhibit prominent relativistic outflows known as jets, whose synchrotron radiation can be detected in the radio domain. The launching, evolution and variable nature of these sources is still not fully understood. We study 3C 84, because its proximity, brightness and the intermittent nature of its jet makes it a good target to investigate these open questions of the AGN phenomena. 3C 84 (optical counterpart: NGC 1275) is a Fanaroff-Riley type I radio galaxy, located in the Perseus cluster at z = 0.0176. Due to its close proximity, 3C 84 has been a favourable target for observations throughout the entire electromagnetic spectrum, especially for ones in the radio domain. Its most recent activity started 2003, when a new component emerged from the core in the form of a restarted parsec-scale jet. This provided a rare opportunity to study the formation and evolution of a jet (see Nagai et al. 2010, 2014, 2017 and Suzuki et al. 2012). The highest resolution results were obtained by Giovannini et al. (2018), who imaged the source with the Global VLBI Network together with the Space Radio Telescope, RadioAstron. This enabled them to capture the limb-brightened structure of the restarted jet and measure its collimation profile from ~350 gravitational radii. In this work I present the 22 GHz RadioAstron observations carried out 3 years later, in a similar configuration, but with a significantly different sampling of the space baselines than the ones presented in Giovannini et al. (2018). The calibration was carried out in the Astronomical Image Processing System (AIPS), whereas imaging was done in Difmap (Shepherd 1997). The aim of this thesis work was to obtain a high-resolution image of the source, measure the collimation profile of the restarted jet, and compare the results with those of Giovannini et al. (2018) and verify the observed source structures and measured jet properties, if possible. Comparing the images of the two epochs (angular resolution of the 2016 observations is 0.217x0.072 mas at Pa=-49.6°), they both show a similar structure, with the radio core, a diffuse emission region (C2), and the hotspot (C3) at the end of the restarted jet. Edge-brightening is confirmed in the jet and the counter-jet. However, the jet has advanced ~1 mas, corresponding to the velocity of 0.55c. C3 has moved from the center of the feature to the jet head, indicating an interaction between the jet and the clumpy external medium (Kino et al. , 2018 and Nagai et al., 2017). The base of the jet has also changed between the observation, approximately by ~20°. In the light that in the 1990s the jet pointed towards C2, then swinged westwards when the jet emerged (Suzuki et al., 2012 and Giovannini et al., 2018), and on the 2016 image has moved towards its initial position. This suggest a precessing jet, observed and modeled by Dominik et al. (2021) and Britzen et al. (2019). Measuring the brightness temperature of the core and the hotspot shows a signifacant drop of 70% and 50% since the 2013 measurements, respectively, due to emission of jet material and the expansion of the jet. Jet width measurements between 1200 and 19000 gravitational radii reveal a less cylindrical collimation profile, with r ~ z0.31 – where z is the de-projected distance from the core and r is the width of the jet. The evolution of the restarted jet’s profile from quasi-cylindrical (Giovannini et al. 2018) to less cylindrical implies that the cocoon surrounding the jet (Savolainen, 2018) cannot confine the jet material as it moves further from the core. The measured collimation profile corresponds to a slowly decreasing density, and more steeply decreasing pressure gradient in the external medium. Since the closest jet width measurement is only at 1200 gravitational radii from the core (here the jet width is 750 gravitational radii), it cannot confirm the wide jet base measured by Giovannini et al. (2018) at 350 gravitational radii. Based on this result, we arrive at the same conclusion as Giovannini et al. (2018), that the jet is either launched from the accretion disk, or it is ergosphere-launched, but undergoes a quick lateral expansion below 1000 gravitational radii.

Core galaxies are bright elliptical galaxies that contain a shallow central surface brightness profile. They are expected to form in mergers of massive gas-poor elliptical galaxies that contain supermas- sive black holes (SMBHs) in their respective centres. During the merger process, these black holes form a coalescing binary, which causes the ejection of stars from the centre of the galaxy merger in complex three-body interactions, resulting in the creation of a low-luminosity core. I have studied whether core galaxies can form according to the formation model described above. I analysed the results of seven galaxy merger simulations done using KETJU, a simulation code specifically made for studying the dynamics of supermassive black holes in galaxies. KETJU is a regularised tree-code, combining both the GADGET-3 tree-code and an AR-CHAIN integrator. This allows for the simultaneous simulation of both general galactic dynamics and accurate particle motion near black holes, respectively. All seven simulations consisted of a merger of two identical galaxies. Six of the simulations had galaxies with equal mass central SMBHs, where the mass of the black holes changed from one simulation to another, and ranged from 8.5 × 10 8 M to 8.5 × 10 9 M . For the sake of comparison, the galaxies in the seventh simulation did not contain SMBHs. The other properties of the merged galaxies were determined in such a way, that the resulting merger remnants would be as similar as possible to the well studied core galaxy NGC 1600. Naturally, these properties were identical across all of the simulation runs. By calculating the surface brightness profiles of the merger remnants in the simulation results, I found out that only simulations that contained SMBHs produced remnants with cores. Furthermore, I identified a clear positive correlation between the size of the core and the mass of the coalescing binary SMBH. Both of these results corroborate the theory, that the cores are formed by interacting SMBH binaries. This interpretation of the results was further enforced by the fact that, according to their velocity anisotropy profiles, stellar orbits near the centre of the remnants were tangentially dominated, implying that stellar particles on more radial orbits had been ejected from the system. I also generated 2D maps of the stellar line-of-sight velocity distributions in the simulated merger remnants. These maps showed kinematic properties similar to observed core galaxies, such as "kinematically distinct cores". Finally, I compared both photometric and kinematic properties of the simulated merger remnant containing the largest SMBH binary to the observed properties of NGC 1600. I found that the simulation and the observations agree well with each other. Since the properties of the simulated merger remnants follow theoretical expectations and is in general good agreement with the obser- vations, I conclude that the formation of the cores in bright elliptical galaxies is likely caused by coalescing binary black holes in dry mergers of elliptical galaxies.

Over the last thirty years more than five thousand exoplanets have been discovered around a wide variety of stellar objects. Most exoplanets have been discovered using the transit method, which relies on observing the periodic brightness changes of stars as their planet transits in front of them. The discovery efficiency of these planets has been strongly enhanced with the advent of space telescopes dedicated to the discovery of planets using the transit method. Planetary signals in the photometric data of active stars can be challenging to find, as the surface features of the stars combined with their rotation might produce signals which are orders of magnitude stronger than those caused by the planetary transit. The question of what statistical methods should be applied to account for the innate variability of stars in order to identify the transits of exoplanets in the lightcurves of active stars is being investigated in this thesis. I test a number of statistical methods in order to combat stellar activity and to identify planetary transit signals. The rotation period of the star is investigated using the Lomb-Scargle and likelihood ratio periodograms. Starspot induced variability is approximated with a number of sinusoids, with periods based on the star's rotation period. Additional stellar activity is filtered out using autoregressive and moving average models. Model fittings are performed with least squares fitting, and using samples generated by the Adaptive Metropolis algorithm. After the lightcurve has been detrended for stellar activity, the likelihoods of planetary transit signals are assessed with a box-fitting algorithm. Models are compared with the Bayesian and Akaike information criteria. Planetary characteristics are then estimated by modeling the shape of the transit lightcurve. These methods are tested and performed on the lightcurve of HD~110082, a highly active young star with one confirmed planetary companion, based on the observations of the TESS space telescope. I find that stellar activity is sufficiently filtered out with a model containing four sinusoid signals. The signal corresponding to the planet is confirmed by the box fitting algorithm, agreeing with results available in scientific literature.

Results from cosmological zoom-in simulations focusing on the formation and evolution of massive early-type galaxies starting from redshift z=50 until the present time are presented in this thesis. In addition, the dynamics of supermassive black hole (SMBH) binaries found in a subset of the galaxies as a result of galaxy mergers are studied. The first 12 zoom-in simulations were run with GADGET-3, an N-body hydrodynamical tree code on the Puhti supercomputer at the Finnish IT Centre for Science (CSC). The zoom-in regions were chosen from a low resolution large-volume simulation, run with the University of Helsinki computing cluster Kale. The SMBH binary dynamics were simulated with the regularized integrator code KETJU. In total three simulations using KETJU were run on the CSC supercomputer Mahti. The GADGET-3 simulations included both dark matter only runs and runs containing also baryons. The simulations were run at two different resolutions. No significant differences between the two resolutions were found for the dark matter only runs, whereas the runs including a baryonic component showed large differences. The medium resolution runs had too low particle numbers and too large particle masses to correctly resolve star formation and feedback, leading to rotation curves that missed their central peaks. The results from the high resolution simulations agreed with earlier published results. The rotation curves peak in the central regions of the galaxies with the curves becoming almost constant at large radii. The star formation rates peak in the redshift range z~2-3 and at smaller redshifts star formation can momentarily increase due to the occuring of galaxy mergers. At the present day, all of the studied galaxies include mainly old stellar populations, resulting in red colours for all the galaxies. Still, the galaxy formation efficiency parameter of each galaxy is somewhat higher than what is seen in the observations. Finally, all three SMBH binaries for which the dynamics were studied using KETJU led to a coalescence of the two black holes. The orbital decay of SMBH binaries occur in three phases: dynamical friction, three-body interactions and gravitational wave emission. The merger times strongly depend on the eccentricity of the binaries with the semimajor axis of the binary with the highest eccentricity decreasing the fastest. This is expected from theory, with the evolution from subparsec scales to coalescence agreeing very well with the theoretical evolution taking into account the post-Newtonian correction term 2.5PN, which is the lowest order post-Newtonian term responsible for gravitational wave emission.

This thesis presents the results from seventeen collisionless merger simulations of massive early-type galaxies in an effort to understand the coalescence of supermassive black holes (SMBHs) in the context of the Final Parsec Problem. A review of the properties of massive early-type galaxies and their SMBHs is presented alongside a discussion on SMBH binary coalescence to motivate the initial conditions used in the simulations. The effects of varying SMBH mass and stellar density profiles in the progenitor initial conditions on SMBH coalescence was investigated. Differing mass resolutions between the stellar particles and the SMBHs for each physical realisation were also tested. The simulations were performed on the supercomputers Puhti and Mahti at CSC, the Finnish IT Centre for Science. SMBH coalescence was found to only occur in mergers involving SMBH binaries of equal mass, with the most rapid coalescence observed in galaxies with a steep density profile. In particular, the eccentricity of the SMBH binary was observed to be crucial for coalescence: all simulations that coalesced displayed an orbital eccentricity in excess of e=0.7 for the majority of the time for which the binary was bound. Simulations of higher mass resolution were found to have an increased number of stellar particles able to positively interact with the SMBH binary to remove orbital energy and angular momentum, driving the binary to coalescence. The gravitational wave emission from an equal mass SMBH binary in the final stages before merging was calculated to be within the detection limits required for measurement by pulsar timing arrays. Mergers between galaxies of unequal mass SMBHs were unable to undergo coalescence irrespective of mass resolution or progenitor density profile, despite the binary in some of these simulations displaying a high orbital eccentricity. It was determined that the stellar particles interacting with the SMBH binary were unable to remove the required orbital energy and angular momentum to bring the SMBHs to within the separation required for efficient gravitational wave emission. A trend between increasing mass resolution and increasing number of stellar particles able to remove energy from the SMBH binary was observed across all the simulation suites. This observation is of paramount importance, as three-body interactions are essential in removing orbital energy and angular momentum from the SMBH binary, thus overcoming the Final Parsec Problem. As such, it is concluded that the Final Parsec Problem is a numerical artefact arising from insufficient mass resolution between the stellar particles and the SMBHs rather than a physical phenomenon.