Original Title:
Study of the earthquake site-effects of the broader Karditsa urban area and deterministicstrong ground motion simulation for the main seismic/active faults of the Thessaly area, using field measurements, stochastic simulations, and Geographic Information Systems
Summary:
The current master's dissertation was submitted in partial fulfillment of the
requirements of Master of Science in Prevention and Management of Natural
Hazards, Faculty of Geology and Geo-environment, National and Kapodestrian
University of Athens. In its first part, this study aimed at estimating the
ground response for the simulated seismic motions from strong earthquakes,
within the broader area of Karditsa (central Greece). For this purpose, the
geotechnical / geophysical background of the city of Karditsa was studied, by
employing existing microzonation data (Lekkas, 1998) and performing ambient
noise (HVSR) measurements. The analysis of microzonation geotechnical data led
to the creation of the Vs30 model for the area, which showed small-scale
spatial variations of the shear wave velocity, Vs, for the uppermost 30m. This
was partly expected due to the known geological setting of the study area. In
the second part, a more generic approach was adopted, concerning a larger study
area (broader area of the whole Thessaly plain). In this approach, it was
attempted to simulate the damage distribution for the 1954, M=7.0 Sofades
earthquake and the 1957, M=6.8 Velestino earthquake, using synthesized seismic
motions, which were appropriately converted to macroseismic intensity (Modified
Mercalli scale, IMM), after incorporating the site-effect through the local
geology (the geological formations of the area have been studied extensively in
the past, Papanikolaou and Sideris, 2007). Since no instrumental recordings
were available for the specific events, macroseismic intensities (IMM up to 9+)
for these events were used, as they have been observed in the broader Thessaly
area. For the simulations, a modified stochastic finite-fault method (EXSIM
algorithm, Motazedian and Atkinson, 2005) was used, in order to reproduce the
damage distribution of both earthquakes.The simulation approach attempts to
combine existing earthquake information and appropriate scaling relations with
surface geology, in order to investigate the efficiency and usefulness of the
available macroseismic data (Papazachos et al., 1997). For the estimation of
the fault dimensions, the calibrating equations for fault geometry from moment
magnitude proposed for dip-slip faults by Papazachos et al. (2004) and
Papazachos et al. (2006) were utilized. In order to account for site-effects on
the observed seismic motions, a new digitized geological map for the broader
Thessaly basin was created, using the 1:50000 geological maps of the Greek
Institute of Geology and Mineral Exploration (IGME). The geological formations
were grouped according to their age and mapped on appropriate soil classes,
according to EC8.Synthetic time series were estimated for different rupture
scenarios. In order to compare the obtained stochastic motions with the
historical (macroseismic) information, various relations between PGA and PGV
(obtained from the stochastic records) and macroseismic intensity were
considered, allowing the generation of synthetic (stochastic) isoseismals. For
each soil class, different site amplification factors were tested, according to
local geology (Skarlatoudis et al., 2003; Klimis et al., 1999). Moreover, a
trial-and-error optimization of each fault position has been performed, using
the available seismological information, as well as the published neotectonic
data for these events and the broader southern Thessaly fault zone
(Papastamatiou and Mouyaris, 1986; Mountrakis et al., 1993).The results confirm
the applicability of the proposed approach. The finally determined positions of
both faults are different than previously proposed, in agreement with the
available neotectonic information. The observed macroseismic intensities are
also in very good agreement with the stochastic simulation predictions,
verifying both the usefulness of the approach, as well as of the macroseismic
data used. Site-effects show an excellent correlation with the geological
classification employed. Moreover, the constant amplification factors of
Skarlatoudis et al. (2003) show a very good agreement with the observed
amplifications, whereas the generic transfer functions proposed by Klimis et
al. (1999) seem to lead to higher site amplifications, not observed in the real
data, probably due to the very large thickness of Quaternary formations in the
broader Thessaly basin. These large thicknesses lead to very low resonant
frequencies, significantly different to the generic transfer functions of
Klimis et al. (1999), as verified by appropriate HVSR measurements performed
along selected profiles in the southern Karditsa area.
