CC BY
15
RUSSIAN JOURNAL OF EARTH SCIENCES, VOL. 8, ES1001, doi:10.2205/2005ES000195, 2006
Macroseismic and instrumental data comprehensive analysis: Earthquake of June 2, 1930 in Catalonia (Spain)
R. Tatevossian
Institute of Physics of the Earth, Russian Academy of Sciences, Moscow, Russia A. Ugalde, and J. Batlio
Ebro Observatory, Horta Alta, 38, 43520 Roquetes, (Tarragona), Spain R. Macia1
Dept. Matematica Aplicada II, Universitat Politecnica de Catalunya, Barcelona, Spain
Received 15 December 2005; revised 20 December 2005; accepted 10 January 2006; published 14 February 2006.
[1] Modern concept of seismic hazard assessment is based on the assumption that what was observed in the past could likely be expected in the future. It could be easily recognized by comparison of spatial distribution of maximum intensity of shaking (Imax) from past earthquakes and actual seismic hazard maps. This makes crucially important the re-evaluation of parameters of earthquakes close to the maximum observed ones in a certain area: even a small change of the parameters can affect considerably the final hazard assessment. A case study of the earthquake on 2 June, 1930 in southern Catalonia, Spain, is presented. A comprehensive analysis of macroseismic and instrumental data leads to a good agreement in magnitude re-evaluation. The earthquake magnitude MS is within the interval 4.5-4.8, which is significantly larger than that reported earlier (3.9). Relatively accurate magnitude evaluation allows us also to assess the hypocenter depth to be at 20-30 km, deeper than it is commonly assumed for this region. Occurrence of large earthquakes in southern Catalonia at various depths (from 10 km to 30 km) reflects the presence of seismogenic structures which are able to produce earthquakes with magnitudes at least as large as 4.5-4.8. Orientation of isoseismals lets us also to suggest that these seismogenic structures are oriented almost perpendicular to strike of topographic elevations, which follow along the seashore. Modern topography does not inherit deep (mid and low crust)
structures. INDEX TERMS: 0935 Exploration Geophysics: Seismic methods; 7223 Seismology: Earthquake interaction, forecasting, and prediction; 7230 Seismology: Seismicity and tectonics; KEYWORDS: Earthquake intensity, Iberian Peninsula, Macroseismic analysis, Old seismograms, Seismic hazard assessment.
Citation: Tatevossian, R., A. Ugalde, J. Batllo, and R. Macia (2006), Macroseismic and instrumental data comprehensive analysis: Earthquake of June 2, 1930 in Catalonia (Spain), Russ. J. Earth. Sci., 8, ES1001, doi:10.2205/2005ES000195.
Introduction
[2] The modern concept of seismic hazard assessment is based on the assumption that what was observed in the past could likely be expected in the future. Usually this principle
1 Also at Laboratori d’Estudis Geofisics “Eduard Fontsere”, Barcelona, Spain
Copyright 2006 by the Russian Journal of Earth Sciences.
ISSN: 1681-1208 (online)
is stated explicitly but even if not so, it is followed implicitly. It could be easily recognized by comparison of the spatial distribution of maximum intensity of shaking (Imax) from past earthquakes with seismic hazard maps. Tatevossian et al. [2006] showed recently that the spatial distribution of Imax (the observed ones supplemented by calculated values based on the epicentral intensities) over the Spanish territory is well correlated with the peak ground acceleration (PGA) contours obtained within the frames of the Global Seismic Hazard Assessment Program [Jimenez and Garaa-Fernandez, 1999]. This fact makes crucially important to
0.5° 0° 0.5° 1° 1.5° 2° 2.5° 3°
Figure 1. Spatial distribution of seismicity in Catalonia and seismic hazard map. Earthquakes in southern Catalonia with magnitude higher than 3.5, are shown in black stars. White star is for a preinstrumental time period earthquake with epicentral intensity I0 = 6.5. Dots are earthquake epicenters (white dots are for historical earthquakes, black ones are for seismic events since 1901). Contours are isolines of PGA (in m sec-2) taken from GSHAP map [Jimenez and Garcia-Fernandez, 1999].
re-evaluate the parameters of earthquakes close to the maximum observed ones in a certain area: even a small change of them can affect considerably the final hazard assessment. In this paper comprehensive analysis of macroseismic and instrumental data for accurate re-evaluation of earthquake parameters is presented. The most promising time period for this objective is the so-called early instrumental period. Before that time period we simply do not have instrumental data, and for the time period after the 1960’s, when the World Wide Standardized Seismograph Network (WWSSN) was set in operation, the instrumental observations ensure relatively high data quality, which is enough to get reliable earthquake parameter determinations.
[3] In this work we will address to the seismicity of southern Catalonia, Spain. Though Catalonia is not large, the spatial distribution of seismicity over its territory is rather inhomogeneous (Figure 1). The highest activity both in historical and modern times is concentrated in its northern part along the Pyrenean Range, approximately up to Barcelona. On the other hand, the part from Tarragona to Ebro valley in the south presents moderate seismic activity. According to the seismic catalogue of the Spanish National
Geographic Institute (IGN) [Martinez-Solares and Mezcua, 2002; Mezcua and Martinez-Solares, 1983], (IGN Database, 2004, http://www.mfom.es/ign/), earthquakes with magnitudes greater than 4 are not known in this part of Catalonia. The spatial distribution of seismicity is also reflected in the hazard map (Figure 1). PGA values along the Pyrenees are greater than or equal to 1.25 m s-2, whereas to the north of Barcelona it exceeds 1.75 m s-2. On the other hand, the isoline of 1.0 m s-2 contours southern Catalonia.
[4] The importance of every earthquake for hazard assessment becomes higher in areas of low and moderate seismicity. At the same time, the parameters of small earthquakes (M < 4) are usually defined with greater uncertainty, especially before the operation of the WWSSN started, because only few stations could record them. In a certain sense, the level of seismic activity justifies our choice of the study region.
[5] Stars in Figure 1 show the relatively large earthquakes occurred in southern Catalonia. Black stars correspond to seismic events with M = 3.5-4 (larger magnitudes are not known in this region). The epicentral intensity (I0) of the 1845 earthquake (white star) is the highest reported ever.
Table 1. Parameters of the earthquakes of 2 June, 1930, and 13 February, 1949 from the IGN catalogue [Mezcua and MarUnez-Solares, 1983].
Date Time Latitude, Longitude, H, mbLg N Io Map Epicentral
(yyyy/mm/dd) (hhmm) °N °E km area
1930/06/02 0152 41.250 1.000 — 3.9 5 5 yes Cornudella
1949/02/13 0506 41.067 0.883 — 4.0 1 5 yes Colldejou
Its magnitude (in fact there were 3 shocks on October 1, 3 and 7) is not defined in the IGN catalogue. The epicen-tral intensity of the first shock is 4, whereas for next two shocks it is 6.5. These are the highest reported I0 in the area. But probably I0 is somewhat exaggerated because of superposition of macroseismic effects from the sequence of shocks.
[6] We will focus our study on the earthquake of 1930. It clearly can be considered as a seismic event of the early instrumental period. Its magnitude in the IGN catalogue is 3.9 whereas I0 is 5 and hypocenter depth is not assigned. It has to be noted that the source depth is neither reported for 1949, nor for 1976 earthquakes with magnitudes 4.0 and 3.6; both with epicentral intensity 5. The absence of hypocenter depth determination could be interpreted as an evidence for low accurate solution for this earthquake. This is again a reason to look more carefully at the seismicity of the region, which was not well monitored instrumentally probably up to mid of 1970’s. The importance of source depth for seismic hazard analysis is evident from the macroseismic field equation [Shebalin, 1971]:
Ii = bM — v log(R2 + H2)1/2 + c, (1)
which in the epicentral area becomes:
Io = bM — v log H + c , (2)
where Ii is the intensity at a distance R km from the epicenter, Io is the epicentral intensity, M is the magnitude, H is the hypocenter depth in km, and b, v, c are coefficients which are equal to 1.5, 3.5 and 3.0 respectively for earthquake sources within the Earth crust. These values of coefficients were proved in most of the world seismic active regions [Kondorskaya and Shebalin, 1982; Shebalin et al., 1974] and they suggest a strong dependence of intensity with depth. It has to be emphasized that the absence of depth for largest earthquakes is an essential drawback not only for hazard assessment but also for the understanding of seismotectonic features governing the seismicity of the region.
[7] Further we will present the record lines for the 1930 earthquake from the parametric catalogue, then we will examine the macroseismic data referenced in the catalogue and finally we will revise the macroseismic and instrumental data to get mutually consistent interpretation.
Reported Parameters for the 1930 Earthquake
[8] We accept as the reference source of information the IGN earthquake catalogue (IGN Database, 2004, http://www.mfom.es/ign/), [MarUnez-Solares and Mezcua, 2002; Mezcua and MarUnez-Solares, 1983], which is the official catalogue of earthquakes in Spain. The parametric lines for the 2 June, 1930 and 13 February, 1949 earthquakes in the IGN catalogue are shown in Table 1. There are two versions of the isoseismal map for the 1930 earthquake in Mezcua [1982], each referring to different sources of information. Both versions are reproduced in Figure 2. The 1949 earthquake map is reproduced in Figure 3 [Meezcua, 1982]. All maps have been re-sized to be approximately in the same scale.
[9] Even a brief examination of the isoseismal maps reveals some incoherencies. Though magnitudes of the 1930 and 1949 earthquakes are almost the same (3.9 and 4.0), the felt area of the first one is larger. It is also much larger (more than 2 times) the area contoured by the intensity IV isoline (regardless to which version of the isoseis-mal map of the 1930 earthquake we compare with the 1949 map). This could hardly be attributed to different attenuations of seismic waves radiated by both earthquakes, because the distance between their epicenters is about 20 km, so the source-locality travel paths are almost the same for both earthquakes. An explanation could be the assumption of a deeper source for the 1930 earthquake. But a deeper source of the 1930 earthquake will become in contradiction with equal magnitudes and epicentral intensities (I0 = 5) for both earthquakes. According to equation (1) and (2), the epicentral intensity, magnitude and felt area of the 1949 earthquake are coherent with the assumption of a relatively shallow source (10-15 km), so the discrepancy between these parameters for the 1930 event has to be solved. For this purpose we have to verify both macroseismic and instrumental data.
Analysis of Macroseismic Data
[10] The intensity data were read from the maps in Figure 2 (a and b, named Source 2 and Source 3, respectively) and they were supplemented with data from two other sources: the archive of the Cartographic Institute of Catalonia [Susa-gna and Goula, 1999, named Source 1] and Fontsere [1940] (named Source 4). All these data are summarized in Table 2.
Figure 2. Two versions of the isoseismal maps for the 2 June, 1930 earthquake [Mezcua, 1982], (a) source SSIS and (b) source Fontsere [1940].
Figure 3. Isoseismal map for the 13 February, 1949 earthquake [Mezcua, 1982]; source Fontsere [1951].
The localities for which the intensity values from 2 or more data sources coincide are highlighted in bold. There are only 9 such cases from a total number of 74 localities. On one hand, this fact demonstrates the lack of detailed and accurate data, which make possible unique intensity assessments and, on the other hand, it proves that different data sources present independent researches. However, in most of the cases the intensity discrepancy does not exceed 1 degree, which could be accepted as a reasonable accuracy of assessment for the majority of reported intensities. Of course in some cases the error could be greater. The last column of the table compiles the intensity degrees considered in this paper. In agreement with European Macroseismic Scale [Granthat, 1998] we follow the convention that sign “-” means uncertainty and not the precision, i.e. I=4-5 means that we are not able to distinguish between 4 or 5 and not that the intensity is 4.5 (which would imply an accuracy of 0.5 degree). The epicenter locations from different sources do not differ significantly because all of them are within a 7 km radii circle. It also has to be mentioned that though none of the sources report observed intensities greater than 5, the Source 1 assesses an epicentral intensity 5-6.
[11] The map of intensity data points considered in this work is presented in Figure 4. Question marks denote doubtful cases, such as when the intensity in a locality differs significantly from the ones for cluster of nearby localities. For example, intensity 4 is assigned to locality Sant Feliu de
Llobregat by Source 1, but there are other four close localities with intensity 2. Other type of doubtful case is when the reported locality is in obvious contradiction with any reasonable attenuation, like in Solsona for which the only source is Source 1. This source assigns intensity 4 to this locality, which is at about 100 km from the epicenter of an earthquake with epicentral intensity 5. The same source assigns an intensity degree 4 to Riumors (42.229°N; 3.044°E), which even falls out of the presented map boundaries, with an epicentral distance of about 180 km. Nevertheless, such strong incoherencies are rare (6 points from a set of 74 localities). This also confirms our assumption that the accuracy for most of the intensity assessments is not worse than 1 degree. It is important to emphasize that the fact of adding the information from two new sources does not change the main features observed when comparing the isoseismal maps of Figures 2 and 3: the felt area and the Ii =4 area for the 1930 earthquake are larger than those for the 1949 seismic event. Moreover, this difference becomes sharper.
[12] The isolines drawn in Figure 4 are calculated using equations (1) and (2). The epicenter location is taken from the IGN catalogue. Because all the earthquake locations from different sources are close to each other, the arbitrariness of this choice does not affect the result. Then we plot 8 isoline sets for different depth values (from 5 km to 40 km with a 5 km step). In the macroseismic field equations a mean radius of isoline is assumed. For the elliptical iso-
6 of 11
Table 2. Intensity data points from different sources and assessed in this paper.
Locality Source 1 Source 2 Source 3 Source 4 Final Locality Source 1 Source 2 Source 3 Source 4 Final
Aiguamurcia Y Y Molins de Rei 2 2
Alcanar 2-3 2-3 Montblanc 4 4
Alcaniz 4-5 4-5 Mont-Roig del Camp 4 4
Aldea. L’ 4 3-4 3-4 Mora d’Ebre IV 3 3-4
Aldover 4 4 Mora la Nova IV 4
Alfarras 3 3 Nulles IV 4-5 4-5
Alio Y Y Penarroya de Tastavins 2-3 2-3
Ametlla de Mar. L’ 5 4 5 4-5 Pobla de Montornes. La III 4 3-4
Amposta 4 3-4 3-4 Pont d’Armentera. El Y Y
Arnes 2-3 2-3 Prat de Llobregat. El N 2
Balaguer IV 4 4 4 Reus Y 4 4 4
Barcelona Y Y Riba. La Y Y
Batea 2-3 2-3 Riudoms 5 4 4
Beceite III 3 3 Riumors IV 4
Bellpuig IV 4 4 Sant Feliu de Llobregat IV 4
Borges Blanques. Les 5 4 4-5 Santa Coloma de Queralt IV 3 4 3-4
Capellades 4-5 4-5 Sarral 4 3 3-4
Cervera IV 2 4 3 Salou IV 4 4 3-4 4
Cornudella de Montsant 5 5 Solivella 3-4 3-4
Cretas 2-3 2-3 Solsona IV 4
Espluga de Francoli. L’ V 5 5 5 Tarragona IV 5 5 4-5
Falset IV 4 4 3 4 Tarrega III 3 3 3
Flix IV 4 5 4-5 Tivissa IV 4 4
Gandesa IV 4 4 3 3-4 Torre de l’Espanyol. La IV 4
Garcia 2-3 2-3 Tortosa V 4 4 3-4 3-4
Godall Y Y Ulldecona IV 4 4 4 4
Horta de Sant Joan 4 4 Valderrobres 3 3
Hospitalet de Llobregat 2 2 Vails IV 4 4 4
Igualada N 2 2 Vendrell. El 3 3
Lilia 4 4 Verdu III-IV 3-4
Lleida IV 4 4 4 Vilafranca del Penedes III-IV 3 3-4
Maella 3-4 3-4 Vilanova i la Geltru IV 3 4 3-4
Martorell N 2 2 Vila-Rodona IV 4 4
Maso. La 3-4 3-4 Vila-Seca IV 4
Mazaleon 3 3 Vilella Alta. La 3 3
Mequinenza 4 2 3 Vinaros 3 3
Miravet 4-5 4-5 Xerta 4 4
Note: Source 1 corresponds to the archive of the Cartographic Institute of Catalonia [Susagna and Goula, 1999]; Sources 2 and 3 are the values reported by Mezcua [1982] in Figures 2a and 2b, respectively. Source 4 is Fontsere [1940]. Column Final of the Table is the intensity degrees assessed in this paper. “N” means, “not felt” and “Y” means “felt”.
ES1001 TATEVOSSIAN ET AL.: MACROSEISMIC AND INSTRUMENTAL DATA COMPREHENSIVE ANALYSIS ES1001
0° 0.5° r 1.5° 2° 2.5
Figure 4. Intensity data points and theoretical isoseismals (H = 30 km, elliptical model) for the 2 June, 1930 earthquake.
lines we assumed that Rmean = (ab)1/2, where a and b are the ellipse semiaxis. The ratio a/b depends on the number of isoline from the epicenter. The ratio is larger for those, which are close to the epicenter. We also vary the orientation of isolines. The isoline set shown in Figure 4 gives the best separation of plotted data (it should be remembered that the symbol “-” is interpreted as uncertainty). The set corresponds to the following parameters: H = 30 km; (a/b)v = 2.5; (a/b)iv = 2.25; (a/b)m = 2.2.
[13] To fit the epicentral intensity I0 = 5 and the isoseis-mal radii in Figure 4 the earthquake magnitude has to be set to Ms = 4.8. Taking into account the accuracy of intensity assessments in the localities together with the number of model variables used for the calculation of each set of isoseismals, one has to be cautious with the obtained magnitude value. However, we have strong reasons to conclude that a larger earthquake magnitude than that reported in
the IGN catalogue is expected. To be more confident we have to verify instrumental data.
Analysis of Instrumental Data
[14] According to Hughes and Bellamy [1935] the 1930 earthquake was recorded on 13 stations at epicentral distances ranging from 0.6° to 11.2°. The fact that the earthquake was recorded up to 1200 km distance already supports a relatively large magnitude value. The reported magnitude by the International Seismological Summary (ISS) Karnik [1969] is Ms = 4.6 (based on data of 2 stations) whereas in the column of remarks a magnitude M = 5.0 calculated by Munuera [1963] with a local magnitude scale is given. Our assessment (Ms = 4.8) based on macroseis-
Figure 5. Example of one digitized signal recorded by the NE-SW component Wiechert seismograph at Toledo Observatory. The best fits for the seismic moment calculation for P and S waves are also shown.
mic data analysis is in a very good agreement with these values. The source depth is not defined in Karnik [1969]; it is only marked as “normal”. On the other hand, the 1949 earthquake is not consigned by the ISS. Karnik [1969] gives a magnitude (4.2) (brackets in the original) taken from Munuera [1963]. However, as it has been shown recently [Lopez and Munoz, 2003] the methodology used by Munuera [1963] presents some shadows really difficult to elucidate now. In an unpublished research, Sanchez-Contador [1988] calculated a local magnitude (Ml) equation for the Mainka’s seismographs and a duration magnitude (Mt) equation for the Vicentini seismograph of the Fabra Observatory (FBR) in Barcelona, situated approximately at one degree of distance from the epicentral zone. The application of these equations gives ML = 5.0 and Mt = 4.9 for the 1930 earthquake and ML = Mt =4.5 for the 1949 event. Samardjieva et al. [1998] also calculated the Mt and Ms magnitudes for the Horizontal Wiechert seismograph at Toledo Observatory (TOL) and obtained 4.5 and 4.1 values, respectively, for the 1930 earthquake. The 1949 event was not recorded at that observatory.
[15] Although all presented data sustain the assumption that the 1949 earthquake had a lower magnitude than the 1930 event, in order to shed light on these discrepancies the instrumental magnitudes for both events were re-calculated using the original seismographic recordings summarized in Table 3. All the recordings have been previously scanned, digitized and processed as described in Dineva et al. [2002]. However, some problems with the stylus inscription, time marks and not damped instruments make most of the records not valid for the seismic moment calculation. The only useful records are those from Toledo Observatory. Thus, the moment magnitude (Mw) has been calculated for the 1930 event. The obtained ground displacement spectra (U) of the digitized recordings have been modelled following Brune [1970, 1971] by fitting:
U (w) = Uo/(1 + (w/wc)Y) (3)
using the non-linear x2 criteria where U0 is the low-frequency level and /c(wc = 2nfc) the corner frequency.
[16] Then, the seismic moment was estimated using the formulation of Keilis-Borok [1960] from the low-frequency
Table 3. List of the original seismograms and recording constants for the 1930 and 1949 events.
Earthquake of 2 June, 1930
Station Type of instrument Component Magnification To(s) Damping
TOL Wiechert NE-SW; NW-SE 480; 420 11.6; 12.1 0.46; 0.47
Wiechert Z 110 4.3 0.38
EBR Mainka N-S 230 14.8 0.29
Mainka E-W 123 7.8 0.33
Vicentini Z 62 0.8 -
Vertical Pendulum N-S 110 2.6 -
FBR Mainka N-S 50 9.8 0.37
Mainka E-W 80 9.9 0.45
Earthquake of 13 February, 1949
Station Type of instrument Component Magnification To(s) Damping
EBR Mainka N-S 187 15.4 0.25
Mainka E-W 220 10.8 0.25
Vertical Pendulum N-S 230 2.5 -
FBR Mainka N-S 64 9.0 0.30
Mainka E-W 73 9.0 0.34
Vicentini Z 125 0.9 —
level of the spectra of body-waves:
Mo = (4npv3Uo)/(G(r)R^C) , (4)
where p = 2.7 g cm-3 is the density in the source region; v the wave velocity; Uo the low-frequency level of spectrum in m-s; G(r) the geometrical spreading factor which depends on the distance r; Rev is the correction for the radiation pattern, which takes an average value of 0.4 for P waves [Wyss and Brune, 1968] and 0.63 for S waves [Boore and Boatwright, 1984]; and C = 2.0 is the correction for the free surface [Moskvina, 1987].
[17] The moment magnitude, Mw, is calculated using the empirical relation [Hanks and Kanamori, 1979]:
2
Mw = 3 log Mo - 6 (5)
with M0 given in N-m. Figure 5 shows one example of the digitized records and the ground displacement spectra fits
(3) for P and S waves. From the obtained Uo values and after applying (4) and (5) results give a value of Mw = 4.4 for P waves and Mw = 4.2 for S waves. Results are also larger than the value previously reported in the IGN catalogue (Ms = 3.9). However, the relationship between both magnitude scales show that Ms values are larger than Mw, for Mw less than about 6 [Utsu, 2002].
[18] On the other hand, although a moment magnitude calculation for the 1949 earthquake has been impossible, the original records give us information about the different size of both events. The Mainka instruments of EBR and FBR stations recorded both events with unchanged working characteristics. Figure 6 shows, in the same scale, the records on
the Mainka N-S component of EBR. The Ebro observatory seismic bulletin states an epicentral distance of 56 km for the 1930 event and 32 km for the 1949 one. A rough visual comparison of maximum amplitudes supports a difference on magnitude between the two events of almost half unit.
Conclusions
[19] There are different concepts of macroseismic intensity. Some seismologists suppose, that there are so many uncontrollable factors influencing the intensity of shaking in a given locality, that it makes no sense to look for any regularity in spatial distribution of intensities. In other words, it doesn’t matter how big is discrepancy between intensities expected from equation (1) and reported in a locality. Especially, if the reported intensity referrers to original questionnaires, which, supposed, automatically ensures high accuracy of intensity assessments. Others assume that, though intensity is a complex phenomenon indeed, anyway, it is not a kind of miracle: mostly source magnitude, depth, and source-locality distance govern spatial distribution of intensities. Local ground effects are very important, too, but they cannot completely alter in general regular character of intensity distribution. The second concept being physically reasonable seems more preferable.
[20] An obvious inconsistency between the reported parameters for the 2 June, 1930 earthquake in southern Catalonia in the IGN catalogue and the already published iso-seismal maps for this seismic event provoked revision of the
Figure 6. Example of the (a) 2 June, 1930 and (b) 13 February, 1949 earthquake records at EBR Observatory (N-S component of the Mainka Seismograph). Both seismograms are plotted at the same scale.
available data. The fact, that instrumental records, being absolutely independent from macroseismic ones, also prove a larger magnitude, than previously evaluated one, support the assumption that intensity of shaking is controlled (at least in general) by regular physical laws.
[21] Comprehensive analysis of macroseismic and instrumental data leads to a remarkably good agreement in magnitude re-evaluation. Results show that the earthquake magnitude Ms is within interval 4.5-4.8, which is significantly
larger than the value previously reported in the IGN catalogue (Ms = 3.9). The relatively accurate magnitude evaluation allows us also to assess the hypocenter depth to be at 20-30 km. On the other hand, the instrumental data analysis gives values of moment magnitudes Mw = 4.4 for P waves and Mw = 4.2 for S waves, also larger than the Ms value previously reported. However, Ms values should be larger than Mw, for Mw less than about 6 [Utsu, 2002]. Finally, results of the isoseismal map analysis of the 13 February, 1949
earthquake are in agreement with the previously reported magnitude and relatively shallow depth (10-15 km). Thus, it can be concluded that the occurrence of large earthquakes in southern Catalonia at various depths (from 10 km to 30 km) reflects presence of seismogenic structures there that are able to produce earthquakes with magnitudes at least as large as 4.5-4.8. The orientation of isoseismals lets us also to suggest that these seismogenic structures are oriented almost perpendicular to strike of elevations, which follow along the seashore. The modern topography does not inherit deep (low crust) structures. The same situation was observed near the western termination of the Caucasus near the Black Sea shore, as revealed by the Lower Kuban, November 9th, 2002 earthquake [Tatevossian et al., 2003]. In the latter case this conclusion was also supported by the fault plane solution: its strike was in remarkable agreement with the elongation of isoseismals.
[22] Acknowledgments. We are very indebted to Jose Manuel Martinez-Solares and M. Aranzazu Izquierdo of the Instituto Geografico Nacional (IGN), Madrid, Spain, for giving us all the macroseismic information of the IGN Database and for making us available digitally the IGN Seismic Catalogue. We also thank to Teresa Susagna, of the Catalonian Cartographic Institute and Fabra Observatory, who provided us with seismograms and macroseismic information on the southern Catalonia earthquakes. This work has been partially supported by a Scientific Program NATO grant, the RFBR grant 04-05-65004, and the MEC Projects HUM2004-04259/HIST and CGL2004-20332-E.
References
Boore, D. M., and J. Boatwright (1984), Average body-wave radiation coefficients, Bull. Seismol. Soc. Am., 74, 1615.
Brune, J. N. (1970), Tectonic stress and the spectra of seismic shear waves from earthquakes, J. Geophys. Res., 75, 4997.
Brune, J. N. (1971), Correction, J. Geophys. Res., 76, 5002.
Dineva, S., J. Batllo, D. Mihaylov, and T. van Eck (2002), Source parameters of four strong earthquakes in Bulgaria and Portugal at the beginning of the 20th century, J. Seismology, 6, 99.
Grunthal, G., (Ed.) (1998), European Macroseismic Scale, in Cahiers du Centre Europeen de Geodynamique et de Seismolo-gie, vol. 15, p. 99, GFZ, Luxembourg.
Fontsere, E. (1940), Mapa de isosistas del terremoto del 2 de Junio de 1930, Publicado de J. Galbis, II, 169.
Fontsere, E. (1951), Los temblores de tierra catalanes de los anos 1948 y 1949, Bol. Sec. Met. Sis. Obs. Fab., V(38, fig. 1), 255.
Hanks, T., and H. Kanamori (1979), A moment magnitude scale, J. Geophys. Res., 84, 2348.
Hughes, J. S., and E. F. Bellamy (1935), International Seis-mological Summary for 1930, 426 pp., University Observatory, Oxford.
Jimenez, M. J., and M. Garcia-Fernandez (1999), Seismic hazard assessment in the Ibero-Magreb region, Annali di Geofisica, 42(6), 1057.
Karnik, V. (1969), Seismicity of the European Area, Part I, edited by D. Reidel, 364 pp., Kluwer, Dordrecht, Holland.
Keilis-Borok, V. I. (1960), Investigation of the mechanism of earthquakes, Sov. Res. Geophys. (in (English trans.)), 4, 29.
Kondorskaya, N. V., and N. V. Shebalin (1982), New catalog of strong earthquakes in the USSR from ancient times through 1977, 608 pp., NOAA, Boulder, USA.
Lopez, C., and D. Munoz (2003), Formulas de magnitud en los boletines y catalogos espa?oles, Fisica de la Tierra, 15, 49.
Martinez-Solares, J. M., and J. Mezcua (2002), Catilogo sismico de la Peninsula Ib!rica (880 a.C.-1900), Monografia, num. 18, p. 253, 1 map., Instituto Geografico Nacional, Madrid.
Mezcua, J. (1982), Catalogo general de isosistas de la
Peninsula Iberica, 262 pp., Instituto Geografico Nacional, Madrid.
Miezcua, J., and J. M. Martinez-Solares (1983), Sismicidad del area Ibero-Mogrebi, 300, 1 map. pp., Instituto Geografico Nacional, Madrid.
Moskvina, A. G. (1987), Spectral response of the earth crust to seismic vibrations, Izv. Akad. Nauk. USSR, Physics of the Earth (in Russian), 23(1), 25.
Munuera, J. M. (1963), Datos basicos para un estudio de
sismicidad en el area de la Peninsula Iberica, vol. XXXII, 93 pp., Memorias del Instituto Geograifico y Catastral, Madrid.
Samardjieva, E., G. Payo, J. I. Badal, and C. Loipez (1998), Catalogue of Digital Historical Seismograms (1912-1962), 315 pp., Wiechert Seismograph, Geophysical Observatory of Toledo, Instituto Geograifico Nacional, Madrid.
Sinchez-Contador, P. (1988), Sismicidad catalana 1907—1987 y modelo de Q-1 en los Pirineos Orientales, Tesina d’Especialitat, Escola Tecnica Superior d’Enginyers de Camins, Canals i Ports, p. 30, Univ. Politecnica, de Catalunya.
Shebalin, N. V. (1971), Notes on predominant periods, spectrum and source of a strong earthquake, in Seismic studies for construction, Problems of engineering seismology, vol. 14, p. 50, Nauka, Moscow.
Shebalin, N. V., V. Karnik, and D. Hadzievski (1974), Catalogue of earthquakes of the Balkan region, I, 600 pp., UNDP-UNESCO Survey of the seismicity of the Balkan region, Skopje.
Susagna, T., and X. Goula (1999), Atles sismic de Catalunya, 608 pp., Institut Cartografic de Catalunya, Barcelona.
Tatevossian, R. E., K. G. Pletnev, A. Yu. Byakov, and V. L. Shestopalov (2003), Lower Kuban earthquake of
November 9, 2002: results of macroseismic survey, Phys. Solid Earth (in Russian), 39(11), 42.
Tatevossian, R., A. Ugalde, J. Aptekman, and A. Petrossian (2006), On the problem of homogeneous presentation of seismic history of large territories, Phys. Solid Earth, 42(2), 37.
Utsu, T. (2002), Relationships between magnitude scales, in International Handbook of Earthquake and Engineering Seismology, Part A, p. 733, Academic Press, London.
Wyss, M., and J. N. Brune (1968), Seismic moment, stress and source dimensions for earthquakes in the California-Nevada region, J. Geophys. Res., 73, 4681.
J. Batllo, A. Ugalde, Ebro Observatory, Horta Alta, 38, 43520 Roquetes (Tarragona), Spain
R. Macia, Dept. Matematica Aplicada II, Universitat Politecnica de Catalunya, Pla de Palau, 18, 08003 Barcelona, Spain
R. Tatevossian, Institute of Physics of the Earth, Russian Academy of Sciences, 10 Bol’shaya Gruzinskaya ul., Moscow, 123995 Russia, (ruben@ifz.ru)
