Browsing by discipline "Geophysics"

The ground motion induced by an earthquake and its attenuation as a function of magnitude and distance can be estimated by using various ground motion prediction equations (GMPE). From a historical perspective, the Fennoscandian shield has been a seismically quiet area with a scarcity of strong earthquakes. This has made the area nominally an area with a low seismic hazard. The lack of quantative strong motion data in an area with low seismic hazard and the subsequent lack of advanced theoretical models of seismic response have hindered the developement of GMPEs for the region. Using more direct empirical methods, which do not depend on pre-existing models and simulations of Fennoscandian seismicity, and by taking advantage of the comparatively large Fennoscandian shield ground motion database two GMPEs targeted for the Fennoscandian shield area were created. The created GMPEs are based on an existing attenuation relationship that targets Eastern North America, which is a stable continental area similar to the Fennoscandian shield. The current and historical stage of relative seismic quietness was preceded by a considerably more active phase during and after the deglaciation of the Weichselian continental ice shelf ca. 9000 - 15000 years ago. There is ample geological evidence that this postglacial faulting caused some of the largest known earthquakes that have ever occurred in a stable continental area. The moment magnitudes of the largest postglacial earthquakes have been estimated to have been in excess of 8.0. The magnitude contrast between the geologically recent postglacial seismicity and the available Fennoscandian ground motion database, which consists mostly of low magnitude events, implies that an empirical method can not provide a GMPE capable of reliably estimating the ground motion induced by a potential large postglacial faulting event. It is unlikely that such an event could occur today, but not completely impossible. The two GMPEs presented here are based on a ground motion database consisting of 6465 recordings from 1701 events with magnitudes between -1.0 – 5.2. The events were observed at 84 seismic stations around the Fennoscandian shield between 2003 and 2014. The first GMPE is an empirical model which takes an existing GMPE and uses a non-linear least squares regression method to refit the constant coefficients of the model to our regional ground motion database. The second GMPE is a referenced empirical model which works by multiplying the ground motion prediction of an existing GMPE with a function of certain seismological parameters. The multiplying function's coefficients are then fitted to the ground motion database. The resulting equations provide a reasonably good model of the peak ground acceleration (PGA) and spectral accelerations (SA) at 9 different frequencies: 0.5 Hz, 1.0 Hz, 2.5 Hz, 5 Hz, 10 Hz, 20 Hz, 25 Hz, 33.3 Hz, and 40 Hz. The GMPEs were further assessed by applying them to an independent regional earthquake and various higher magnitude external events that have originated in a presumably similar stable continental area. Based on these comparisons, the upper magnitude limits of applicability were independently estimated for PGA and each SA frequency. A tendency of the ground motion estimate to improve with increasing frequency at higher magnitudes can be seen. The distance ranges of the GMPEs were determined to be identical with the original base model.

Finland is situated in an intraplate area of low seismicity. Seismic hazard analyses require an assessment of regional maximum earthquake magnitude. One of the methods for estimating maximum magnitude is relating it to the dimensions of active faults. Intraplate earthquakes usually occur when pre-existing zones of weakness are reactivated in response to the ambient stress field. Because minor earthquakes rarely cause surface ruptures, the reactivated faults have to be studied by indirect means. In this study structural lineaments are used as proxies for old shear zones, faults and fractures. Their orientation with respect to the crustal stress field is determined in order to find potentially unstable faults. Firstly the orientation of the stress field is determined by reviewing literature and available data on crustal stress in Finland. The main force causing the compressive stress field in Fennoscandia is the spreading of the mid-Atlantic ridge, which is why the plate motion of Finland relative to North America is also taken into account. An estimate of 115° to 135° is reached for the azimuth of the maximum horizontal stress in Finland. The stress regime is mostly reverse (minimum principal stress is vertical) according to stress indicators and focal mechanisms. The lineaments are split into straight segments for azimuth calculation. The segments are then divided into optimal orientation categories based on the horizontal angle between the segment and the maximum horizontal stress. Reverse faulting takes place perpendicular to and normal as well as transfer faulting takes place parallel to the maximum horizontal stress. The direction of strike-slip faulting depends on the coefficient of internal friction, which is around 0.6 for solid rock and as low as 0.2-0.4 for pre-existing fractures. With these values the Coulomb failure criterion gives an optimal angle of 30° to 40° to the maximum horizontal stress for strike-slip faulting. The lineament segments with the different faulting categories are shown with different colours in order to better visualise the regions hosting similar faulting directions. Some coefficients of internal friction were also calculated based on available stress magnitudes by assuming frictional equilibrium of pre-existing, optimally oriented zones of weakness. The data were scarce and only available for shallow depths. The calculated coefficients are quite high (0.7-0.8) near the surface and decrease with depth down to 0.4. A maximum earthquake magnitude of approximately 7 is suggested based on lengths of the lineament segments. An earthquake of such magnitude has a very low probability of occurring in Finland. This leads to the conclusion that the datasets used are too coarse for reliable estimates of fault length. Finally, the lineament orientations are compared with seismicity data. In western Lapland earthquakes are clearly linked to a reverse and transfer faulting system. In order to find differences in earthquake source mechanisms the earthquakes are divided into three depth categories. It seems that shallow earthquakes of depths less than 5 km most often occur near lineaments likely to reactivate as reverse and transfer faults, whereas earthquakes deeper than 15 km occur closer to lineaments optimal for strike-slip faulting. Earthquakes between the depths of 5 km and 15 km occur near lineaments of all orientations. The lineament orientation and seismicity maps will hopefully prove useful in further studies concerning the present structural framework of the Finnish crust. For reliable estimates of maximum earthquake magnitude based on fault length in Finland, faults should be studied in greater detail.

In this thesis I study the radiation balance and heat budget of a multiyear sea ice floe drifting in the central Arctic ocean. The objectives of the study were to quantify the vertical partitioning of shortwave- and longwave radiation and to quantify the different components of the heat budget of the floe in question, both inside and at its interfaces. The measurements were set up at 88 26.6N, 176 59.88W on 8th of August and carried out for ten days. The measurements were made as a part of the fourth Chinese National Arctic Expedition CHINARE2010. The measurement setup consisted of a net radiometer, four PAR-sensors, a pyrano-albedometer, three spectral radiometers, daily snow pit measurements, weather observations and six ice corings. With the data from these studies I was able to quantify the rate of melting and fluxes of heat both at the surface and at the bottom of the ice. The data allowed for examining the fraction of transmitted and conducted heat but were insufficient for properly quantifying the internal changes and spectral composition of the shortwave radiation at different depths. The surface was observed to be losing heat mainly in the longwave part of the spectrum. The average net radiation on top of the ice on wavelengths between 200 nanometers and 100 micrometers over the period was -25.0 Watts per square meter. The heat fluxes of the shortwave and longwave radiations were of opposite directions and the negative heat flux of the longwave radiation dominated until a distinct change in the radiative conditions on 17th of August. For the remainder of the period these heat fluxes nearly balanced each other and the average net radiation was -2.1 Watts per square meter. The latent and sensible heat fluxes were observed to have a minor role in the surface heat budget with averages of -1.5 Watts per square meter and -0.03 Watts per square meter respectively. The ice was observed to melt primarily at the bottom at a rate of 0.5 cm per day driven by the input of heat from the underlying ocean. Melting at the surface was not apparent until before the last two days of studies, when the upper layer of the snow cover melted. The changes in sea ice and snow cover were visually observed to exhibit significant spatial variability even on a single floe.

Yksi tapa tarkastella maankamarassa etenevien seismisten aaltojen vaimenemista on määrittää tutkittavalle alueelle seisminen Q-arvo, joka kuvaa maa- tai kallioperän keskimääräisiä vaimenemisominaisuuksia. Seismisellä Q-arvolla on myös taajuusriippuvuus, eli sillä on eri arvo eri taajuuksilla. Tässä työssä määritettiin taajuusriippuvat Q-arvot P- ja S-aallolle tutkimalla näiden ensisaapujien alusta alkavan kolmen sekunnin aikaikkunan maksimiamplitudeja. Kohdealue tutkimuksessa oli Pohjois-Suomi. Havaintoaineisto koostui Suomen seismiseen asemaverkkoon kuuluvien, viiden seismografiaseman maanjäristysrekisteröinneistä. Asemat sijaitsivat Hetan, Sodankylän, Oulun, Kuusamon ja Tornion alueilla ja maanjäristykset olivat tapahtuneet enintään 200 km:n episentrietäisyydellä. Eri asemien rekisteröintejä tutkittiin yhteensä 181 kpl. Tutkimuksessa käytetty menetelmä on nimeltään coda-normalisointimenetelmä, jossa P- ja S-aaltojen amplitudiarvot normalisoidaan coda-aaltojen amplitudiarvoilla. Eri hyposentrietäisyyksillä saaduista arvoista saadaan regression avulla määritettyä Q-arvo eri taajuuksille, joista linearisoinnin avulla saadaan määritettyä taajuusriippuva Q-arvo. Tässä työssä P-aallolle saatiin taajuusriippuva Q-arvo Qp(f ) = 183(±12)f^0.64(±0.15) sekä S-aallolle Qs(f )=288(±30)f^0.67 (±0.23 ) . Saatuja tuloksia verrattiin samalla menetelmällä eri puolilta maapalloa saatuihin tuloksiin. Pohjois-Suomen alueella S-aaltojen vaimeneminen oli kaikkein vähäisintä verrattuna tektonisesti aktiivisempiin tai geologisesti nuorempiin alueisiin. Tuloksia verrattiin myös sekä Fennoskandian alueelle että geologisesti samankaltaisille alueille määritettyihin Q-arvoihin. Tässä työssä määritetty Qp-arvo ja sen taajuusriippuvuus ovat hyvin samankaltaisia Norjan alueelle määritettyjen Qp-arvojen kanssa. Qs-arvo puolestaan osuu keskivaiheille muihin arvoihin vertailtaessa, ollen hyvin samankaltainen kuin Pohjois-Amerikkaan ja Australiaan määritetyt Qs-arvot. Coda-normalisointimenetelmä osoittautui riittämättömäksi alueelle, jolta on vähän ja epätasaisesti jakautunutta havaintoaineistoa maanjäristyksistä. Seismisten aaltojen vaimenemiseen liittyviä tutkimuksia suositellaan jatkettavan, mutta jollain toisella menetelmällä. Eräs mahdollinen tutkimuskohde tulevaisuudessa voisi olla geometrisen leviämistekijän ja sen etäisyysriippuvuden määrittäminen Suomen kallioperälle.