|Geophysical Journal International juill 2004
The 1928 Plovdiv seequence (Bulgaria): fault model constrained from geodetic data and surface breaks
Dimitrov D.S.1, J.-B. Chabalier2, J.-C. Ruegg2, R. Armijo2, B. Meyer2 and E. Botev3
1 Central Laboratory of Geodesy (CLG), Bulgarian Academy of Sciences, Sofia, Bulgaria.
2 Institut de Physique du Globe, CNRS UMR 7578 et 7580, Paris, France.
3 Geophysical Institute, Bulgarian Academy of Sciences, Sofia, Bulgaria.
Two destructive earthquakes (M=6.8 and M=7.0) occurred on April, 14 and April, 18 1928, near the city of Plovdiv, in Southern Bulgaria. These earthquakes are among the most important that occurred in Europe and the Eastern Mediterranean area during the 20th century. A detailed analysis of the tectonic observations available in the litterature allows us to determine that the 1928 sequence reactivated successively two minor antithetic faults of respectively 38 and 53 km length with mainly normal displacement, consistent with available seismological data. We use an exceptional geodetic data set, formed by 500km long levelling lines and GPS remeasurements of a former triangulation network to characterize the coseismic displacement field. The deformation is asymetric, with major faulting during the 2nd shock, along a north-dipping fault with a maximum of 2m of vertical displacement. We reproduce this displacement field and the surface breaks using dislocations in an homogeneous elastic half-space. This requires for the first shock a south-dipping fault with uniform slip of 0.7 m. The 2nd shock accomodates about 75% of the seismic moment of the sequence on the north-dipping fault with steeper dip near the surface than at depth, and a maximum slip of 2,6 m. Surprisingly these two major earthquakes reactivated minor faults and did not rupture the larger fault in this area. This poses the tectonic and mechanical problems of the mode of deformation in the North Aegean Region.
Key words: sequence of earthquakes, geodetic data, North Aegean, fault model,
On April 14, and 18, 1928, a sequence of two major earthquakes struck the Southern Bulgaria (Richter 1958). Their epicentral area is located near the city of Plovdiv, in the Maritza valley between the Rhodope and the Balkans Mountain belts. This sequence is one of the most important that occurred in Eastern Mediterranean region during the twentieth century. Their magnitude have been estimated to be respectively 6.8 and 7.0 (Karnik, 1969). The two shocks and subsequent aftershocks hit a large area of about 3000 km2 and caused very important damages: destruction of 5 towns and more than 240 villages. Many reports have been published describing the effects of this sequence (Bonchev and Bakalov, 1928; Zelkov, 1929; Mihailovic, 1928a, 1928b; DIPOZE, 1931; Kirov, 1935). However no tectonic analysis of this documents have been done to determine unequivocal tectonic ruptures associated with the two shocks. Before and after the earthquake the National Cartographic Institute of Bulgaria (NCI) performed levelling surveys of the area (Mirkov, 1932; Yankov, 1938; 1945). This 500 km long levelling network is composed by 357 benchmarks that intersect the rupture at 8 points. These data provide an exceptional and unique set of coseismic elevation changes of historical earthquake.
A former triangulation network remeasured by GPS also provide the coseismic horizontal displacements associated with this sequence (Dimitrov and Ruegg, 1994). This extensive set of data is the opportunity to produce a refine fault model of this sequence of earthquake and to understand its role in the processes of deformation tha take place in the North Aegean region that exhibit low rate of deformation (e.g. McClusky et al. 2000; Kahle et al. 2000). A first attempt to explain the geodetic data have been done by Dimitrov and Ruegg (1994), proposing a first order model. In this paper, we have compiled the available tectonic, seismologic and geodetic data to propose a refined fault model of this major sequence. We first revise the available descriptions of the effects of earthquakes (most of them in Bulgarian language) to discriminate the main surface rupture of the two shocks from secondary effects, in relationship with active faults in the epicentral area. We combine this information with available seismological data to determine the fault model of the sequence. Using the levelling data and GPS measurements we propose an improved fault model associated with the sequence. Finally we discuss the larger-scale tectonic implications in relationship with the nearby extensional strains in the North Aegean region.
REGIONAL TECTONICS AND ACTIVE FAULTING
The earthquakes of April, 14 and April, 18 1928 occurred north of the Aegean Sea in the Upper Thrace Lowland in Bulgaria (high Maritza Valley) between the Rhodope massif in the South and the crystalline Balkan Mountain in the North (Fig.1). The structural trend is almost NS in the Dinarides and the Hellenides in the east, turns NW-SE in the Rhodope to EW in the Balkan, as outlined by the topography and the courses of large rivers (Fig. 1). These belts have been formed mainly from late Mesosoic to early Neogene as crustal shortening was taking place between the converging Africa and Eurasia plates (e.g. Dewey et al. 1973). From the post-oligocene the region appears mostly under NS crustal extension and is considered to be the northernmost part of the Aegean stretched domain (e.g. Jackson and Mc Kenzie 1988). The driving mechanism of this extension remains controversial. Some have proposed high stretching rates with crustal thinning during the last 5 Ma (Jackson and Mc Kenzie, 1988) or during a polyphased evolution, during the last 25 Ma (e.g. Jolivet et al. 1994; Burchfield et al. 2000). Another view has emphasized a two stages evolution resulting from recent propagation of the North Anatolian fault into the already slowly extending Aegean domain (Armijo et al. 1996, 1999). In Bulgaria, the normal faults strike mostly EW to ESE and are slightly oblique to the fabric in the Rhodope and the Balkan belts. They are less developed than in central Greece and western Turkey (Fig. 1), where some of them have ruptured during destructive earthquakes with clear surface breaks: Thesaloniki, Ms=6.4 in 1978 (Mercier et al. 1979) and Grevena Ms=6.6 in 1995 (Meyer et al. 1996). In Bulgaria, a sequence of two destructive earthquakes (two shocks separated by 23 minutes), with an estimated magnitude of 7.3 and 7.8 (Karnik 1968) occured in 1904 in the Struma valley (west of the Rhodope, Fig. 1) and have some similitude with the Plovdiv sequence. The high Maritza valley where the 1928 sequence occurred corresponds to a large EW asymetric graben filled by Neogene continental sediments to alluvial Quaternary sediments. A conspicuous normal fault forming the northern border of the Rhodope massif (the Asenovgrad fault, Fig. 1 and 2) lies close to the epicentral area. Ten destructive earthquakes (I>VIII) have been mentioned during the last millenium (UNESCO, 1974) in this area.
OBSERVATIONS OF SURFACE BREAKS AND SEISMOLOGICAL CONSTRAINS
Important surface effects are associated with the 1928 sequence. Breaks of 105 km total length has been reported, producing a large subsiding area in the Maritza valley, with the formation of several temporary lakes and landslides. The descriptions, maps and reports of tectonic observations, have been made soon after the earthquake sequence (Bonchev and Bakalov, 1928; Mihailovic, 1928 a, b; Zelkov, 1929). The most detailed descriptions of surface ruptures, including many photographs, have been published by the National Commission for the mitigation of seismic damages (DIPOZE, 1931) and in two monographs published by the Bulgarian Academy of Science (Mihailovic, 1933; Kirov, 1935). We have reconsidered all these available data to distinguish the main surface breaks from secondary effects for each mainshock.
On Figure 2 is reported the higher intensity contour (I=X) of the isoseismal map published by (Bonchev and Bakalov (1928) and DIPOZE (1931). The respective effects of the two shocks can be clearely identified: the isoseismal map of the April, 14 shock is centred on the city of Chirpan whereas this of the April, 18 shock is 30 km to the west, centred on the city of Duvanja. In the following section we discuss the surface observations together with the available seismological data and the geomorphology.
The April 14 event, activated two zones. The clearest one is associated with the most severely damaged area, located North of Chirpan. It consists of a continuous rupture of 38-km length between Trakia and Orizovo, striking N95° and dipping to the South. Continuous normal scarps with an average of 40 cm down-to-the-south normal slip have been measured, reaching a maximum of 50 cm in middle of the rupture NW of Chirpan (DIPOZE, 1931). The villages distributed along the mapped rupture, Chirpan, Tcherna Gora, Orizovo, Partizanin, Gorno Belovo, Svoboda have been totally destroyed (DIPOZE, 1931). This clearely indicate that this 38 km continuous rupture is associated with a fault activated during the first shock. This rupture outlines the base of a small topographic relief of 10 to 20 m, that may correspond to cumulative slip of a pre-existing minor normal fault that will be called Chirpan fault hearafter. Evidences of dextral rung fissures and horizontal offsets of roads of a few centimetres indicate a small right lateral component of slip (DIPOZE, 1931).
The second activated area of the April, 14 event is located 12-15 km south of the main rupture. Surface breaks occurred on a discontinuous 62-km long segment that follows the southern bank of Maritza River between Nadejda and Belozem. It consists mainly of fissures and small scarps with up to 20-cm down-to-the-north offsets. Numerous extraordinary manifestations have been reported in that area including landslides near Maritza River, craters, flooding. A vertical offset of 3.5 m probably associated to a localised landslide have been reported by Mihailovic (1928). The village of Liubenovo (now Parvomai) have been flooded by 0.7 m of water. By contrast to the northern rupture, the villages located along the southern rupture have not been severely affected by the earthquake (for example Debar located 3 km south of Parvomai, Popovitza and Tchechnigirovo; DIPOZE, 1931). These observations suggest that the southern area has been mostly affected by superficial effects of non tectonic origin as slumping and landsliding of the Maritza River banks. South of the Maritza river bed, a topographic relief of about 20m follows the cracks and slumps. If this scarp has been build by repeated earthquakes, it indicates that the 1928 sequence did not activate significantly this minor fault.
Table 1 lists the six available epicentral estimations of the first shock. Except A6 instrumental determination which is located 100 km away from the area, the others are involved within the two identified ruptures. Determinations A1 to A4 (including instrumental determination A2) are concentrated on an area of less than 10 km centred on the city of Chirpan. In Figure 2 we represent the hypocenter as the barycentre of the solutions (42.18°N; 25.32°E, black star in Fig. 2). The hypocenter of this event is located around 10 km depth. No focal mechanism have been proposed for the April 14 shock except Jackson and McKenzie (1988) from surface ruptures. To determine the focal mechanism of this event, we collect the available data of P first arrivals from European seismic stations (Fig. 2). Position of dilatations and compressions is consistent with normal faulting mechanism. However no reliable mechanism can be determine with these data.
New surface ruptures are associated to the April 18 main shock. The ruptures has been followed for 53-km length with clear scarps down-to the North. From the East, the main rupture started 10 km SW of Parvomai and paralleled the scarp forming the Rhodope massif a direction approximately N118°. It crossed Maritza river between Popovitza and Duvandja and forming an important waterfall of 1.5 m height, followed the Striama river farther NW, with a direction N 144°, and ended around village Dalgo Pole, 20 km north of Plovdiv. A rupture of 10 km length have also been mapped along the Maritza River, connecting the main rupture at the confluence with Striama River (Figure 2). The most important displacements are associated to the eastern part of the April 18 ruptures. Southwest of Popovitza, 1.50 m of vertical displacement has been observed (Bonchev and Bakalov, 1928) that coincide with a significant 30-40m topographic scarp that can be followed up to Parvomai, corresponding to a minor preexisting normal fault. However, no cumulative fault scarps are associated with the north-western extension of this rupture along the Striama rivers. Moreover, many complex secondary fissures and ruptures quasi parallel to the main rupture have been mentioned by the authors in the Maritza-Striama basin. Between Popovitza and Duvandja a vertical offset about 3 m have been observed. The villages between Parvomai and Sodovo have been flooded (Mihailovic, 1933). Evidences of right-lateral components of slip can also be done by the analysis of 62 photographs: en echelon fissures, bayonet features, deformed railway, chimney and monuments (DIPOZE, 1931: Kirov, 1935 ; Rijikova, 1996). The April 18 shock activated unequivocally 53 km of a minor northwest dipping normal fault, that will be called Popovitza fault herafter.
Six estimations of epicentral location have been proposed (Table 1). Except B6 located in the Rhodope mountains the others are concentrated around Popovitza village where the maximum of vertical offset have been observed. We consider only B1 and B2 solutions, located north of the rupture as epicenter (42,18°N 25.05°E (white star on Figure 2). Hypocenter of this event seems to be deeper than the April 14 shock (around 15 km). Two slightly different focal mechanisms of the April 18 shock have been proposed by Glavcheva (1984) and Van Eck and Stoyanov (1996). Dimitrov and Ruegg (1994) calculated a solution very similar to Glavcheva (1984). In this study we collect new data from European stations (27 first arrivals) and determine a reliable focal mechanism presented in Fig. 2 (strike= 300°, dip= 67°, rake = -124°). The plane dipping to the north fits very well with the direction of the main surface rupture. The small right lateral component of slip is consistent with the field evidences.
After the two earthquakes the authors indicate flow variations of 350 mineral springs that can be related to co-seismic strain changes (Muir-Wood and King, 1993). Particular field effects or surface ruptures were reported after the strongest aftershock occurred on April, 25 (M=5.7) near the city of Galabovo, at the eastern end of the rupture and fifteen aftershocks with magnitude M>5 recorded in the following month.
The 1928 Plovdiv sequence reveals reactivation of the two anthetic Chirpan and Popovitza normal faults that correspond to preexisting minor faults. In particular no rupture has been observed along the Asenovgrad fault, 10 km south of Plovdiv, that shows clear morphological evidences for young Quaternary slip with a topographic scarp greater than 1000m (Figure2).
(Discussion sur les magnitudes à terminer avec Dimi). The Magnitude of the two main shocks correspond to the surface magnitude scales with 98 determination for April, 14 and 108 determination for April, 18 earthquakes. Fifteen aftershocks with magnitude M>5 were recorded in the following month. The epicentral intensities Io are assessed according to macro-seismic scale MSK (Christoskov, 2000).
The National Cartographic Institute of Bulgaria (NCI) installed the National levelling network in the Plovdiv region from 1923 to 1927 (Fig.3). The levelling surveys has been performed by classical 1st and 2nd order method of double run, using spirit levels and invar rods. After the seismic sequence of April, 1928, 506 km of levelling lines has been re-measured by NCI in 1929-1930 using the same methodology and instruments. The random error for these surveys is estimated as +/-1.5 mm/km. After adjustment of the two epochs data, Mirkov (1932) published the co-seismic elevation changes, and selected profiles and deduced an accuracy better than 1cm for these data. From the same set of data, Yankov (1938, 1945) proposed an attempt of interpretation. As Dimitrov and Ruegg (1994) we model the same set of data, completed with the Nova Zagora-Maritza levelling line, re-measured in 1940 and the Plovdiv-Karlovo levelling lines re-measured in 1959 (Fig. 3). Finally, we adjust the position of all the available data together, to determine the co-seismic elevation changes of 355 levelling benchmarks (Annexe 1). Figure 3 presents the position of the benchmarks used in this study and Figure 4 some selected profiles. The main results are summarized in the following (Figure 3 and 4).
For the first shock, a 15 km wide asymmetric zone of subsidence is located north of the Maritza River, bordered in the north by the south dipping Chirpan fault, with a maximal down-throw of 48 cm (profiles e). An uplift is located NE of the Chirpan city with a maximal displacement of 24 cm (profiles, e and f). The small displacements associated to the southern rupture of April 14 (levelling lines Parvomai-Dimitrovgrad, Fig. 3, and profile e, f) corroborate that the observed effects along Maritza river are most probably due to local slumps and are not associated to slip on a fault. A large subsidence zone is associated to the main surface rupture associated with April, 18 shock, with a maximal down-throw of 103 cm near Belozem (profile b). An uplift zone located SW of the Popovitza fault reach 53 cm near Popovitza city (profile a, d). The 8 points where can be directly measured the surface vertical displacements (points A to H in Fig. 3 and Table 3) are in good agreement with the map of rupture determined in the previous section. The very small displacements measured on the far levelling lines (Nova Zagora and Karlovo lines) give good confidence of this set of data. In particular no displacement is associated to the Asenovgrad fault (Figure 3 and profile e). Both surface break observations and levelling data indicate that the sequence of Plovdiv earthquakes activated two main faults: the southern dipping Chirpan fault during the first shock and the NNE dipping Popovitza fault during the 2nd shock . These two antithetic faults are not face to face but clearly shifted laterally. The deformation is asymmetric with major faulting of the Popovitza fault (maximum displacement of 1.5m) than for the Chirpan fault (maximum displacement of 0.5m).
The 1st and 2nd order pillar triangulation network has been established and measured in 1925-1926 by NCI. The random error for the angular observations was estimated better than 1.0 arc sec (Mirkov, 1932). In 1958 only 5 triangulation points were re-observed in this region. The maximal angular change of about 11.13'' for a distance of 11.5 km indicates clear deformation of the network with displacements of about 50 cm. Using the well preserved original monuments of the network, we realise GPS measurements in order to determine horizontal co-seismic deformations. The Central Laboratory of Geodesy of Bulgarian Academy of Sciences carried out the survey in 1993 in collaboration with Institute de Physique du Globe, Paris and Cartographic Military Institute of Bulgaria (Dimitrov and Ruegg, 1994). The 1993 GPS measurements being more accurate than the triangulation coordinates, we determine the displacement vectors 1926-1993 using least-square adjustment of the original angle observations, taking as reference the 1993 GPS coordinates. The fixed points is located on Sredna Gora Mountain, far from the affected zone (point 104 in Fig 3 and 5). The results of the adjustment are presented on Table 4 and Fig 5. Despite the crude accuracy of the former triangulation measurements, the displacement vectors are consistent with NS to NNE-SSW extension with maximal displacement of about 83 cm +-20 cm between Popovitza and Parvomai.
To model the coseismic surface displacements we use a dislocation embedded in an elastic half space (Okada, 1985). We proceed with trial and error, integrating the tectonic and seismologic constraints and increasing progressively the complexity of the model to fit the levelling and GPS data. We first define the strike and length of the two main dislocations using the ruptures and the levelling data. For Chirpan fault the 3 points F, G and H (Fig. 3) where levelling lines intersect the rupture allow to define the strike of the dislocation (strike= N95°E, length= 36 km) in good agreement with the mapped rupture. For Popovitza dislocation we determine the strike of the plane by adjustment of the 5 aligned levelling points (A to E in Fig. 3) which separate the hanging wall from footwall (N299°E and length= 31 km). It is consistent with focal mechanism determined previously (N300°E). This direction does not follow the mapped ruptures. We attribute this discrepancy to interactions of the rupture with the river banks at the confluence of the two rivers and secondary effects in the basin (local slumps and landslidings). Secondly we use these two single dislocations to determine the dip of the faults taking a rake fixed to 90° corrsponding to purely normal faulting. Subsidence to uplift ratio that can be calculated at the point A to H (Fig. 3, Table 3) are used to fixe the mean dip at the different intersection points. The best fit of this ratio indicate dips that vary from 40° to 70° (7th column of Table3). For the Chirpan fault, a dislocation dipping at 60°, with 70 cm of slip between the surface and 10 km depth explains satisfactorily the data. Higher complexity is not required by the data. For Popovitza fault the best mean dip varies from one profile to the other (70° for measurement A, D, E and 40° for B and C in Table 3). Moreover no uniform dip allows to fit the hangingwall together with the footwall sides as well as the width of the subsiding area (i.e. on profile b, Fig 4). The particular important uplift of the footwall visible on profiles a and d (Fig. 4) clearly indicate that this fault has a steeper dip near the surface than in depth. Therefore a change of dip with depth is required by the data as proposed previously (Dimitrov and Ruegg 1994). We take 75° dip from the surface to 6.3 km depth and 45° down to 14 km. The variation of the vertical offset along this fault also requires a variation of slip. Therefore we divide the dislocations in 5 subfaults and inverse the levelling data to determine the slip and the depth of the 10 resulting patches. The final model formed by 11 patches is presented in figure 5 (parameters in Table 5) and the fit between data and model on cross section in figure 4.
This model explain well the amplitude, local vertical gradients of displacements and width of the subsided areas. Especially, the model reproduce very accurately the uplift of the footwalls, as seen in the profiles of figure 4. The misfit that could be observed in the subsiding area can be explain by secondary effects in the sedimentary basin. For the eastern extension of Chirpan fault, we maximize the length of the dislocation taking into account the observed surface rupture. The levelling data can accept a dislocation 10 km shorter at the east end, that reduce the seismic moment of this fault to 25%. Levelling data does not require any slip on the antithetic fault of April 14 event along the Maritza river. Taking a north dipping fault that follows the ruptures south of Maritza river, leads to a 30-50 km length dislocation with less than 50 cm of slip between 3 and 4.5 km as proposed by Dimitrov and Ruegg (1994) in a preliminary model. Such a model is not reasonable and inconsistent with length to slip scale laws. All these results corroborate our analysis of documents and photographs that suspects local slumping along the Maritza river bed.
In Figure 5 we plot the horizontal vectors predicted by the model. Except for point 110, the modelled vectors are inside the error ellipsis, indicating good agreement between observations and model. We tried to improve the horizontal fit varying the rake of the slip vector with a right lateral component consistent with tectonic and seismological observations. Taking a rake of 100-120° does not improve the horizontal fit, and has no significant effect on the vertical displacement. These geodetic data can not resolve the lateral component of slip in a range of 30°.
DISCUSSION AND Conclusions
Our model is consistent with available seismologic and tectonic data. Taking a shear modulus of 33 Gpa, the models yields a moment of 1.0 x1019 Nm (Mw=6.7) for the april 14 event and 2.8 x1019 Nm (Mw=7.0) for the April 18 event (details in table 5) consistent with seismological and macro-seismological estimations (respectively M=6.8 and 7.0). For the first shock geodetic data do not resolve the eastern extension of the slip. Taking a dislocation of 10 km shorter in the east leads to a moment of 0.7 x1019 Nm (Mw=6.6) slightly smaller than the estimations. About 73 to 80 % of the total moment is released by the second event indicating asymetric graben system with most active north dipping Popovitza fault (Fig. 6). The antithetic Chirpan fault accommodate only 70cm of slip. It can be interpreted as a consequence of the flexure of the hanging wall due to major faulting observed along the southern north dipping system. The dominance of the southern system being observed during the 1928 sequence (Popovitza fault) and also in the topography (Asenovgrad fault, Figure 2).
The regional seismic activity between 1931 and 1980 in the Maritza valley is moderate with M<5 (Rizhikova and Toteva, 2000) but the precision of localisation does not allow any interpretation. Figure 6 presents the 1980 to 2002 instrumental seismicity (M> 2.5), beginning 56 years after the sequence. Most of the seismicity is located between 5 and 15 km depth with two clusters one located around 1928 sequence and the other in the south around the Asenovgrad fault. Two events with M> 4 have been localized, one in the Sredna Gora and the second near the April 14 epicenter. No significant seismicity can be noticed north of Maritza valley, where a major active fault cut longitudinally the Balkan belt (Fig. 1). The southern cluster occured in 2002 and have been felt by the population of the region of Plovdiv. This seismicity is probably associated with the Asenovgrad fault that bounds the northern border of Rhodope.
The 1928 sequence reactivated a minor normal fault system (cumulative topographic relief < 20-40m), with two earthquakes of magnitude 6.8 and 7. Generally the large normal faulting earthquakes (M>6) that struck the Aegean and the western Turkey in the last 30 years have occured on steep fault with clear morphology recorded by the topography. It is therefore possible that the Plovdiv sequence occured on a large hidden old fault with very low long term slip rate and infrequent earthquakes. However the proximity of the Asenovgrad fault (less than 20km to the Popovitza fault), that exhibits typical normal fault authorize alternative interpretation. The 1928 eathquakes reactivate two subdued faults, part of a much larger fault system with interconnecting segments including the Asenovgrad fault. The rupture on the Popovitza fault mimics the geometry of the Asenovgrad fault farther south. This suggests that the Asenovgrad fault is possibly encountering mechanical obstacles to slip, and that the 1928 faults has developped more recently by migration to the northeast. This interpretation imply that the Popovitza fault is in a progress of growing. This kind of geometry have also been observed for Grevena earthquake (M=6.6, 1995, Meyer et al. 1996). This earthquake ruptured minor segments (topographic relief < 50m) interconnecting two major active quaternary faults farther in the south, that has not been activated during the earthquake. To explain the migration of the deformation, Meyer et al. (1995) suggest that a range, formed by an ophiolitic complex, may have produced a mechanical obstacle to the growing of the major faults. In the Plovdiv region no conspicuous geological structure can explain this migration. We can propose that the 1928 ruptures correspond to the growth of the Asenovgrad fault system to the east, reactivating former structures to explain the large magnitude of the sequence. The main rupture observed along the Popovitza fault follows clearly the right bank of both Maritza and Striama river. The 20-40m cumulative fault scarp observed around Popovitza can be followed up to Parvomai and Dimitrovgrad, where cracks and slumps have been observed during the first shock, with no significant displacement. It can indicate long term tectonic control of the course of the river, at the southern border of the half graben due to major faulting of the southern faults as shown by 1928 sequence (Figure 6b).
Conséquences sur le mode de déformation dans la région nord egeene avec des systèmes de chaînon et bassin qui construisent le topo et accommode une petite partie de la déformation NS ans le Rhodope.
This work is part of a collaboration between the Bulgarian Academy of Sciences, Sofia and the CNRS-Institute de Physique du Globe, Paris, France.
IPGP contribution no
Armijo R., Meyer, B., King, G.C.P., Rigo, A. & Papaanastassio, D. (1996) Quaternary evolution of the Corinth Rift and its implications for the eveolution of Aegean, Geophys. J. Int., 126, 11-53.
Armijo R, Meyer, B., Hubert, A. & Barka A. (1999) Westwards Propagation on the North Anatolin Fault into the Northern Aegean: Timing and kinematics, Geology, 27, 267-270.
Bonchev S. and P. Bakalov (1928) Les seismes du Sud de Bulgarie, Revue de Societe geologique de Bulgarie, 1, No 2, pp. 8- 21, Sofia. (in French and in Bulgarian).
Burke W.F. and G. Ugurtas (1974) Seismic interpretation of Thrace basin, Proc. Second. Petrolieum Congress of Turkey, TPAO internal report, Ankara,. pp. 229-249.
Christoskov L. and E. Grigorova (1968) Energetic and space-time characteristics of the destructive earthquake in Bulgaria after 1900, Bull. d’Institut de Geophysique, XII, pp. 79-107 (in Bulgarian).
Christoskov L. (2000) Energy end source parameters of the strong bulgarian earthquqkes after 1900, Reports on geodesy, Politechnika Warszawska, No 3 (48), 15-20.
Dimitrov D. and J-C. Ruegg (1994) The 1928 Bulgarian earthquakes: fault geometry from geodetic data and modelling, 1st International Symposium on Deformations in Turkey, Istanbul, pp. 921-932.
Foose R.M. and F. Manheim (1975) Geology of Bulgaria: a review, The American Petroleum Geol. Bull., 59, No 2, pp. 303-335.
Glavcheva R. (1984) Characteristic of the destructive earthquake of April 18, 1928 M=7.0) in Southern Bulgaria, Geophys. Journal, XVI, No 4, pp. 38-44, BAS, Sofia (in russian).
Grigorova E., W. Sponheuer and R. Maaz (1970) Seismicity of Bulgaria. Pure and Applied Geophysics, vol. 81, IV, pp. 101-111.
Jackson J. and D. Mc Kenzie (1983) The geometrical evolution of normal fault systems, Journal of Structural Geology, vol.5, No5, pp. 471-482.
Jackson J. A. and Mc Kenzie D. P. (1988) The relationship between plate motions and seismic tensors, and the rate of active deformation in the Mediterranean and Middle East, geophys. J. Int., 93, 45-73.
Jolivet, L., Brun, J.P., Gauthier, P. Lallemant, S. & Patriat, M., 1994. 3D-kinematrics of extension in the Aege n region from the early Miocene to the present, insights from the ductile crust, Bull. Soc. Geol. France, 165, 195-209.
Jankov K. (1938) Niveauanderungen im Schuttergebiet der Erdbeben in Sudbulgarien am 14. Und 18. April , Zeitschr., fur Geophys., Jahrgà. 14, Heft 1/ 2, pp. 20-26.
Karnik V. (1969) Seismicity of the European Area, parts 1, Ed. Academia, Prague.
Kirov K. (1935) Study of the earthquakes of April 14 and April 18, 1928, Monography, BAS, Sofia, (in Bulgarian), 116 pp.
Kirov K. and E. Grigorova (1961) Seismicity in the Maritza valley, Bull. Geophys. Ins., No 2, BAS, Sofia, pp. 5-55 (in Bulgarian)
McKenzie D.P. (1972) Active tectonics of the Mediterranean region, Geophys.J.R.Ast.Soc., 30, pp. 109-185.
Mercier, J.L., Mouyaris, N., Simeakis, C., Roundoyanni, T. & Angelidhis C. (1979) Intraplate tectonics: a quantitative study of the faults activiated by the 1978 Thessaloniki earthquakes, Nature, 278, 45-48.
Meyer B., R. Armojo & D. Dimitrov (2002) Active faulting in SW Bulgaria: possible surfase rupture of the 1904 Struma earthquakes, Geophys. J. Int. 148, 246-255.
Mihaillovic J. (1928) Les grands tremblements de terre en Bulgarie en 1928, CRAS, Paris, t. 186, pp. 1562-1563.
Mihailovic J. (1928) Tremblements de terre de Bulgarie en 1928: Situation geologique des regions devastees et dislocations diverses, CRAS, Paris, t.186, pp.1741 -1743.
Mihailovic J. (1933) “Trusovi u jujnoi bugarskoi”, Posebna izdania, kn.XCV, kn. 24, Beograd, 283 p. (in Serbe).
Mihailovic J. (1933) La seismicite de la Bulgarie du Sud, UGGI Monography, serie B, Fasc. 3, Ed. Slovo, 54 p. Beograd (in French).
Mirkov M. (1932) Precision levelling measurements in the seismogenic zone of southern Bulgaria, Bull. Inst. Nat. Geogr., pp. 34-39, Sofia. (in Bulgarian).
Muir-Wood R. and G.C.P. King (1993) Hydrological signatures of earthquake strain, J. Geophys. Res. Vol.98, B12, pp. 22035-22060.
Okada Y. (1985) Surface deformation due to shear and tensile faults in a half-space, Bull. Seismol.Soc. Am., 75, 4, pp. 1135-1154.
DIPOZE (1931) Direction Report for the mitigation of the 1928 bulgarian earthquakes, Monography, 421 p., Sofia, (in Bulgarian).
Richter C. F. (1958) Elementary seismology, Ed. W. H. Freeman, San Francisco, 768 p.
Rijikova S. (1996) Earthquake catastrophes, Sofia, 192 p. (in Bulgarian).
Rizhikova S. and Toteva T. (2000), Maritsa region seismic activity during the fifty years “calm” period (1931–1980), Reports on geodesy, Politechnika Warszawska, No3 (48), 51-56.
Scholz C.H. (1990) The mechanics of earthquakes and faulting, Ed. Cambridge Univ. Press, New York, 439 pp.
Tapponnier P. (1977) Evolution tectonique du systeme alpin en Mediterranee, Bull. Soc. Geol. France, XIX, No 3, pp. 437-460.
UNESCO (1974) Catalogue of Earthquakes, Survey of Seismicity of Balkan Region, Skopje, Part I, II, 600 p.
Van Eck T. and T. Stoyanov (1996) Seismic hazard modelling for southern Bulgaria, Tectonophysics, 262, pp. 77-100.
Yankov K. (1945) Changes in ground level produced by the earthquakes of April,14 and April,18 1928 in southern Bulgaria. In: Tremblements de Terre en Bulgarie, No 29-31, pp. 131-136, National Metheo Ins., Sofia, (in Bulgarian).
Zelkov Y., (1929) Report, Department scientific, Agricultural Ministry, Sofia, 10, Vol. 1 and 2., pp. 3-6, (in Bulgarian).