DISS 3.2.GreDaSS: Seismogenic Source GRIS051 - Rymnio Fault

Source Info Summary

Commentary

Pictures

References

COMMENTS Rymnio Fault is the middle segment of a larger fault zone, the Aliakmonas CS (GRCS050). The fault was partially reactivated along with the Palaeochori Fault segment (GRGG050) during the May 13, 1995 Kozani earthquake. An old strike slip antithetic structure, the Chromio Fault (GRGG053), was also reactivated as a dip-slip fault according to spatial aftershock distribution (Hatzfeld et al., 1997; 1998; Resor et al., 2005) and field observations (Pavlides et al., 1995; Mountrakis et al., 1998; Doutsos and Koukouvelas, 1998). Although many studies refer to the co-seismic results of the event, very few things were known before. The Rymnio Fault has a subdued morphology and the least prominent fault scarps. According to morphotectonic mapping and field observations before and after the event (Pavlides et al., 1995; Mountrakis et al., 1998; Doutsos and Koukouvelas, 1998; Meyer et al., 1996), few fault scarps are visible on surface in contrast with the prominent easternmost Servia segment (GRGG052) which however was not reactivated at all. A geometric segment barrier between the Rymnio and Servia Faults that prevented the rupture propagation producing a right stepping geometry is proposed by Pavlides et al. (1995) and Mountrakis et al. (1998). The coseismic ground ruptures followed most of the few fault scarps filling occasionally the gaps between the branches. Strike, length and rake are based on morphotectonic mapping and field observations from Pavlides et al. (1995) and Mountrakis et al. (1998). Maximum depth is inferred from spatial aftershock distribution of the May 13, 1995 after Hatzfeld et al. (1997; 1998). It is largely accepted that the dip of the fault plane is relatively steep between surface and 9 km depth and progressively decreases in its deeper part (Mountrakis, et al., 1998; Meyer, et al., 1996; Drakatos, et al., 1998). The average dip is inferred from geological data (Pavlides et al., 1995; Mountrakis et al., 1998) and the aftershock spatial distribution (Hatzfeld et al., 1997; 1998). Slip per event and Mw are based on analytical and empirical relationships (Kanamori and Anderson, 1975; Hanks and Kanamori, 1979, Wells and Coppersmith, 1994). OPEN QUESTIONS During the 1995 earthquake, the geometric segment barrier prevented the rupture propagation easternwards to the Servia Fault. Nevertheless, it is unknown how the barrier will behave on a future reactivation. It is also noteworthy that some authors who used InSAR models and seismologic data suggest a single smaller fault despite the observed co-seismic ground ruptures. SUMMARIES Mountrakis et al. (1995; 1996; 1998) According to the authors, the total length of the May 13, 1995 reactivated fault segment is about 30 km long overall, it dips steeper at the surface than in depth and it is separated from the non-activated Servia fault segment by a geometrical seismic segment barrier, near the village of Goules, which causes an angle between them. According to the authors, the total length of the fault segment is about 30 km long overall, the dipping is steeper at the surface than it is in depth and there is a geometrical seismic segment barrier near the village of Goules causing a right-stepping structure with a small change in strike separating the fault from the non-activated Servia Fault segment. Additionaly, a series of sub-parallel antithetic surface fractures mainly striking E-W or ENE-WSW and dipping to the south, were produced during the earthquake. These co-seismic ruptures follow older strike-slip faults but behave as normal ones antithetic to the main seismic fault. The most important of them is the Chromio Fault which was partly activated as a secondary structure. The authors divide the main activated fault into two segments: The Palaeochori and Rymnio segments. They suggest that the Rymnio Fault is the possible starting point of the 1995 rupture propagation after seismological instrumental data from Papazachos et al. (1996; 1997) and geological field observations. During their field study, they observed that the Rymnio segment affects Mesozoic limestones as well as it also shows basinwards migration affecting Neogene, Quaternary and recent deposits. It formed steep scarp slopes, fresh slickensides and fault scree. Several fault planes were additionally observed, probably secondary structures, parallel to the main Rymnio fault segment affecting the Plio-Pleistocene clastic sediments and the recent fault scree forming parallel elongated scarps. The authors did not find clear field evidence for the southwestward extension of the fault to the Vourinos Mountain area, but they identify the fault lineament towards the Palaeochori Fault segment via satellite image processing. Regarding the ground deformations of the 13 May 1995 event, the authors recorded many ground fractures and fissures near the vicinity of Rymnio village, with vertical displacements varying from 1 to 10 cm having occasionally small dextral strike slip component, where in some cases the displacement reached 20 cm in addition with gravitational effects. Pavlides et al. (1995) The authors investigated the broader epicentral area after the 1995, Kozani earthquake. They suggest that the fault is part of a larger tectonic structure forming the Aliakmon River valley. They detected this structure from satellite images as a long and continuous lineament with a total length greater than 80 km. Near to the epicentral area two main segments were recognised, the Palaeochori-Sarakina and Servia Faults. The first is the reactivated fault of the 1995 event which forms a characteristic scarp within the molassic sediments in contrast with the second, which was not reactivated, forming however the most prominent active neotectonic structure bounding the southern shore of the artificial Polyfytos Lake. The two faults are separated by an abrupt change in fault trend causing an angular change between them. The two faults are separated by an abrupt change in fault trend causing an angular change between them. This seismic segment barrier also constrained the aftershock distribution. Concerning the seismogenic fault, the authors suggest a total surface length of about 30 km and a NE-SW strike with a NW dipping. The large amount of co-seismic ruptures was mainly aligned along the main seismogenic fault (N 70°) and along an antithetic fault (N 100°). At the eastern part of the total seismogenic fault trace extension the co-seismic ruptures and the liquefaction phenomena in Rymnio village extend to a length of 2-3 km, while at the central and western parts, where the affected formations are ophiolites, molassic and recent sediments, the total length of the ruptures is about 15 km. The surface ruptures observed at the northern side of the major fault are in branches and comprise a circa 20 km-long secondary antithetic fault structure. The maximum displacements observed in the broader area are on the order of few tens of centimetres. Papazachos et al. (1998) The authors propose a single blind fault plane model in an active regional tensile stress field that caused normal faulting in the central part of the aftershock zone and normal with some right-lateral slip component at the two ends due to the fault intersection with other NE striking structures. The suggested fault parameters from the mainshock’s focal mechanism and the aftershock distribution are: strike 240°, dip 45°, rake -101°, length 30 km, and width 10 km. The mean displacement of the fault plane was calculated to 0.5 m by using analytical and empirical relationships. The authors proposed the following rupture model for the 1995 earthquake: there was pre-seismic aseismic slip in the shallow central area of the fault that induced additional tectonic stress, causing the breakage of barriers and the generation of foreshocks (with magnitude up to 4.5) close to the western boundary of the aseismic slip area. The aseismic deformation was probably accelerated after the generation of foreshocks, inducing high tectonic stress which in few minutes reached the mainshock's barrier ultimate strength and resulted in the generation of the mainshock close to the deepest and eastern boundary of the aseismic deforming area. The generation of the mainshock caused a redistribution of stresses and their concentration, mainly at barriers of the western part of the fault where aftershocks occurred. The rupture initiated at the depth of the focus of the mainshock (-12 km), propagated up-dip and bilaterally and terminated at a depth of about 4 km. They also suggest that water from Polyfyto artificial lake intruded to the faults surface from the eastern end facilitating the fault's motion. Lekidis and Theodoulidis (1995) They studied the strong motion of the 1995 event from the recorded data of an analogue accelerometer that is placed in the building of Kozani's prefecture. The maximum horizontal acceleration, parallel and transversal to instrument's components, was 0.21 g and 0.15 g respectively, the maximum vertical was 0.08 g, and the effective acceleration of the horizontal motion was 0.14 g. Dziewonski et al. (1996) According to the Global CMT Catalogue the following data are available for the May 13, 1995 (Mw = 6.5) earthquake: Depth= 15 km, Scalar Moment = 7.64E+25. The fault plane suggested by the focal mechanism solution has 243° strike, 47° dip and -97° rake. Meyer et al. (1996; 1998) According to the InSAR results, the seismic fault of the 13 May 1995 earthquake is constrained more or less to the Palaeochori Fault (GRGG050) and its antithetic structure occupying a small part of the Rymnio Fault, although some unruptured fault scarps are observed between the Palaeochori and Servia Faults. Chiarabba and Selvaggi (1997) In this paper the authors produced earthquake tomographies by inverting P wave arrival times from 656 aftershocks of the 1995 event in order to resolve the velocity structure of the upper 12 km of crust. They consider the Rymnio Fault the southwestern continuation of an en echelon structure formed by the Servia (GRGG052) and Deskati (12 km to the SW) Faults. Earthquake tomographies show the Servia Fault plane dipping 60° NW at depth and 80° near the surface. Hatzfeld et al. (1997; 1998) The authors plotted the aftershock spatial distribution with which they constrain the seismic fault of the 13 May 1995 event mainly to the Palaeochori Fault (GRGG050). However, the existence of some aftershocks in the hanging-wall area of Rymnio Fault implies the partial activation of Rymnio Fault. The focal mechanism computed by body-waveform modelling indicates a normal fault striking N240°, dipping 40° and having the centroid depth at 11 km. The total scalar moment is calculated to 6.2 E18 Nm. Papazachos and Papazachou (1997; 2003) The authors summarise the 13 May 1995 earthquake effects. According to them many villages in Grevena and Kozani area were destroyed where the macroseismic intensity reached IX+ in MM scale. Along an ENE-oriented 30 km-long rupture zone, ground fissures with up to 15 cm displacement were observed and the maximum damage occurred. The earthquake was preceded by foreshocks. Ambraseys and Jackson (1998) The authors try to associate faults with historical and recent earthquakes. Among others there is also the 1995 event for which they provide the following information: Ms= 6.5, strike= 240°, fault’s mechanism= normal, length= 15km, maximum observed lateral and vertical displacement= 0 and 5 cm, respectively. About the fault’s characteristics they propose an explicit surface faulting with discontinuous tracing and moderate evidence of location. Chatzipetros et al. (1998) The authors carried out a palaeoseismological investigation of the adjacent Palaeochori Fault (GRGG050). After excavating 4 trenches, three prior to the 1995 events were discovered and collected samples were dated using thermoluminescence and C14. Three surface faulting paleoevents were identified at ca. 8.97, 36.7 and 72.5 ka BP (TL dates). This means that the 1995 earthquake seismogenic fault was not active during historical times, since there is no displacement of layers containing pottery or burned wood bearing layers. Furthermore, it seems that the fault was inactive during late Holocene. Hence, the 1995 earthquake was an out of sequence event, because the elapse time since the last major event is 8.97 ka instead of 30. The estimated average recurrence interval for large magnitude (Ms> 6.0) earthquakes on this fault is between 10,000 and 30,000 years. Doutsos and Koukouvelas (1998) According to the authors, the Sarakina seismic fault, reactivated during the May 13, 1995 Kozani earthquake is geologically divided into three second order segments. The Palaeochori central segment is overlapped to the south by the Nisi and underlapped to the north by the Rymnio segments forming an en echelon structure. Coseismic slip along the surface rupture is diminished westwards, from 20 cm in the Rymnio, 10 cm in the Palaeochori, to zero in the Nisi segment. Slip on the Rymnio segment is confined to a well defined range front, whereas slip in the Palaeochori and Nisi segments is transferred into a 3 km-wide zone of diffuse fractures. Concerning the Rymnio Fault, a length of 10 km is suggested and a subsidence of the hanging-wall that formed a Pliocene-Quaternary sedimentary wedge whose thickness exceeds 600 m. Taking into account the 900 m-high fault escarpment across the 45° dipping fault surface, a displacement of 1500 m is estimated. Based on a 1700 m of cumulative displacement across the Sarakina segmented fault which occurred since the last 6 Ma, a long-term slip-rate of about 0.3 mm/a is estimated. Assuming co-seismic maximum displacement of 0.5 m the authors calculate the occurrence of 3500 events during the same time period and a mean recurrence interval in the order of 2000 years. Based on the coseismic slip decreasing and fault length shortening on both three segments, a progressive southwestward younging of the fault activity is suggested. Drakatos et al. (1998) According to the authors, the seismogenic fault plane is steeper near the surface (60-70°) and becomes more horizontal (30-40°) at a depth of ca 10-12 km. Studying the correlation between earthquakes and impending reservoir, they suggest that the 13 May 1995 Kozani earthquake is an event in the framework of the regional seismicity rather than an event triggered by the load of the artificial Polyphyto Lake. Ambraseys (1999) In this paper, the author refers to historical earthquakes prior to the May 13, 1995 event, in the broader area of Kozani. He estimates the produced moment magnitude equal to 7.6 exp18 Nm. Kiratzi (1999) The author reconstructs the mean stress field within the seismogenic volume applying a stress tensor inversion from the Kozani-Grevena 1995 seismic sequence and obtaining a N159°-oriented sub-horizontal sigma3. She estimates a mean dip of about 52°, though the fault plane is steeper between surface and 9 km-depth and shallower in its deeper part. She also points out the existence of a south dipping antithetic fault with plane dipping about 43°. The small amount of sinistral strike-slip motion observed along NW-SE trending structures is inherited from previous deformations. Goldsworthy and Jackson (2000) The authors carried out a morphotectonic investigation about the Rymnio-Servia Fault system which consists of two right-stepping segments, the Rymnio and Servia Faults, each at least 5 km long and both forming steep escarpments more than 500 m high in limestone. The May 13, 1995 Kozani earthquake is not related with any of the two segments but only with the Palaeochori Fault. A morphotectonic observation which supports segmentation is that drainage from the footwalls enters the basin at topographic lows associated with the ends of fault segments or en echelon steps between them. Another observation that supports recent fault activity is that the hanging-wall contains young lake sediments, formed in the fault-bounded basin and tilted to the SE. From field observations, polished fault plane exposures are seen at various points along the Servia segment but are less clear on the Rymnio segment. Fountoulis et al. (2002) After establishing a GPS network after the 13 May 1995 event, the authors suggest that in the broader area the present time deformation process is not only due to seismic activity but also due to aseismic (creep). Kiratzi and Louvari (2003) The suggested parameters from the focal mechanism determined by waveform modelling for the May 13, 1995 earthquake are: depth= 12 km (+1/-2), Mw= 6.4, strike= 242°, dip= 38° and rake= 269°. Vannucci and Gasperini (2003) According to the database of Earthquake Mechanisms for European Area (EMMA) the following data are available for the May 13, 1995 (Mw = 6.4) earthquake: Depth= 12 km, Scalar Moment = 4.898E+18. The fault plane suggested by the focal mechanism solution has 242° strike, 38° dip and -91° slip.