Active tectonics of the Adriatic Region

zyxwvutsrq zyxwvuts zyxwvuts zyxwv Geophys. J. R. asfr. SOC. (1987) 91, 937-983 Active tectonics of the Adriatic Region Helen Anderson* and James Jackson Bullard Laboratories. Madingley Rise, Madingley Road, Cambridge C B 3 OEZ Accepted 1987 May 6. Received 1087 May 6; in original form 1986 May 28 Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 Summary. Seismicity and fault-plane solutions show that the active deformation in the Adriatic region is very varied. West of Messina, N-S shortening occurs with slip vectors representative of the overall AfricaEurasia motion. Along the length of peninsular Italy, NE-SW extension on normal faults is the dominant style of deformation, but changes t o N-S shortening in N. Italy. Inland central and northern Yugoslavia is deforming on strike-slip and thrust faults, and an intense belt of NE-SW shortening continues south along the coast from central Yugoslavia into Albania. South of Albania the shortening in coastal regions is in a more easterly direction. The most remarkable feature of the region is the low level of seismicity in the Adriatic Sea itself, compared with the intense activity in the hgh topographic belts that border it on the SW, NW and NE. The relatively rigid behaviour of the Adriatic allows its motion relative t o Eurasia to be described by rotation about a pole in N. Italy. Anticlockwise rotation about this pole accounts, in a general way, for the change in style and orientation of the deformation in the circum-Adriatic belts. Historical and recent seismicity account for approximately equal rates of extension in central Italy and shortening in southern Yugoslavia of about 2 mm yr-' ; however, these are uncertain by at least a factor of two, and are anyway likely t o be underestimates of the true motion, because of the unknown contribution of aseisniic creep. The Adriatic region resembles, in some ways. other relatively stable continental blocks, such as Central lran and the Tarim Basin, that are caught up within the distributed deformation of the Alpine-Himalayan Belt. The Adriatic, however, is bounded on three sides by the relatively stable Eurasia plate. Its boundary with the African plate is short and ill-defined by seismicity, but is likely t o be located in the Southern Adriatic, near the Strait of Otranto. The present day seismicity shows that the Adriatic, although once perhaps zyx 'Present addrew: D.S.I.R. Geophysics Division. P.O. Box 1320, Wellington. New Zealand 938 zyxwvuts zyxwvutsr zy zy H. Anderson and J. Jackson ‘a promontory of Africa’, is n o longer behaving in this way, and the motions on its boundaries d o n o t directly reflect the Africa-Eurasia convergence. Key words: seismicity, fault-plane solutions, active tectonics, Adriatic 1 Introduction Figure 1. Seismiclty of the western Alpme-Himalayan seismlc belt as reported by the USGS from 1961 t o 1983 August. Earthquakes with reported depthsgreater than 50 k m are not Included. Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 I n this study we use recent and historical seismicity, fault-plane solutions, and young tectonic structures to investigate the active deformation of the Adriatic region. This study was prompted by the need to update an earlier account of the seismotectonics of the western Mediterranean (McKenzie 1972) by the addition of fifteen years of seismicity. McKenzie (1972) recognized that some features of the present deformation between the continental masses of Africa and Eurasia could be described in terms of the relative motion between several small. relatively rigd plates. He tentatively suggested that the Adriatic area is part of the African plate, but noted that there were, at that time, too few fault-plane solutions from large earthquakes t o propose any tectonic interpretation with confidence. Since McKenzie’s study, several earthquakes have occurred that are large enough for reliable fault-plane solutions to be determined. These new fault-plane solutions required a re- zyxw zyx zyxw zyxwvu Active tectonics of the Adriatic region 939 2 Promontory or microplate? Since the first systematic geological studies of the Mediterranean, attention has been drawn to the curved nature of the mountain chain surrounding the Adriatic Sea. This chain of mountains runs through the backbone of Italy as the Apennines, curves tightly around the Po Valley as the Alps, and continues along the Yugoslavian and Albanian coasts as the Dinarides and Hellenides. In contrast with these highly deformed mountainous regions, flat areas like the Adriatic sea-floor and Apulia appear to be structurally simple, This observation led Argand (1924), among others, to suggest that the stable Adriatic area acted as a promontory of the African continent, which has been pushed into the Eurasian continent. This idea gained wide acceptance but has been challenged relatively recently by the suggestion that the stable Adriatic region is a 'microplate' or 'microcontinent' that has acted independently of continental Africa (e.g. Celet 1977). Adria was the name first given by Suess (1883) to a previously emergent area in the position of the present Adriatic Sea. The term is used here following Channell, D'Argenio & Horvath (1979) to refer t o the relatively stable Adriatic area (Po Valley, Adriatic Sea and Apulia), which is surrounded by wide mountain belts (Apennines, Alps, Dinarides, Hellenides) that mark its boundaries. Most workers appear to agree that the Jurassic and Cretaceous complexity of the Adriatic orogenic belts is best explained by a model in which Adria was then a promontory of Africa (Channell et al, 1979; D'Argenio, Horvath & Channell 1980; D'Argenio & Horvath 1984). That Adria continues to act as a promontory is disputed by many authors (Vandenberg & Zijderveld 1982; Celet 1977; Giese & Reutter 1978; Hsu 1982; Morelli 1984) who base their objections on interpretations of palaeomagnetism, crustal structure and recent tectonic style. These authors envisage Adria as an independent 'microplate'. If the focal mechanisms of large earthquakes in the Adriatic area are consistent with those observed in other areas where Africa and Eurasia are in direct contact, then it is Iikely that Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 interpretation and a new kinematic description of the active deformation in the Adriatic region. This paper will not address the driving forces responsible for the observed motions. The Adriatic region is part of the zone of distributed deformation between the African and Eurasian plates. The seismicity within this zone is diffuse (Fig. l ) , but west of Sicily, the largest earthquakes occur within a somewhat narrower zone that extends through N. Africa and Gibraltar to the Azores Triple Junction. The instantaneous Africa-Eurasia pole of rotation defined by the slip vectors of these large earthquakes is located at 27.59"N, 19.74"W (Anderson 1985), which is similar to the pole positions found by Chase (1978: 29.2"N 23.5"W) and Mirister & Jordan (1978: 25.2"N 21.2"W) using longer term data, including sea-floor spreading rates and uansform-fault trends. Earthquake mechanisms change from normal and strike-slip faulting in the Azores region, through strike slip west of Gibraltar to thrust faulting in Sicily. This situation is summarized in Fig. 2. East of Sicily large earthquakes occur in a diffuse zone that includes Italy, Yugoslavia and Greece. Particularly intense seismicity marks the Hellenic subduction zone and the normal faulting in the Aegean Sea (McKenzie 1978). Closer examination of the distribution of shallow earthquakes in Italy, Yugoslavia, Albania and western Greece shows that the seismicity is concentrated in land areas and that few large earthquakes occur within the area covered by the Adriatic Sea. This observation suggests that the deformation in the areas surrounding the Adriatic Sea results from the motion of this relatively aseismic, presumably relatively rigid, Adriatic block. The object of this study is to examine this seismicity in detail. H. Anderson and J. Jackson zyxwvutsrq zyxwvutsrq zyxwvutsrq 94 0 zyxwvutsr Figure 2. Slip vectors calculated for various positions along the Africa-Eurasia plate boundary. Each slip vector is calculated at the position of the centre of the The length of the arrow is proportional to the magnitude of the velocity at each point. arrow for an Africa-Eurasia pole of rotation located at 27.59"N. 19.74"W. The scale arrow is not intended to show any relative motion. Shaded areas indicate the areas of most intense seismicity. Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 Active tectonics of the Adriatic region zy zy zyx 94 1 Adria has been, and still is, a promontory of Africa. If, on the other hand, the deformation is inconsistent with the predicted overall motion between Africa and Eurasia, then Adria could be seen as an independent microplate. The seismicity cannot, however, exclude the possibility that Adria has only recently become detached from the African continent and that it acted as a promontory in the past. Fig. 3 shows that earthquakes are concentrated in a belt that runs through the backbone 47'N Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 4 50 zy 430 zyxwvutsrqponm 4 1' 390 37' 35"N 1( 1 12" I I 16" I 18" I 2 on '"E Figure 3. Seismicity of the Adriatic region, reported as shallower than 50 k m by the USGS from 1963 to 1984 April. All reported events are shown, including those too small o r too poorly recorded for a magnitude t o be determined. 31 14" 942 zyxwvuts zy H. Anderson and J. Jackson of Italy following the Apennine trend and then becomes more diffuse in the Alpine area. Intense activity marks the Albanian and southern Yugoslavian coastal regions. The Po Valley, northern and southern Adriatic basins and Apulia regions are notably less seismic. The seismicity therefore defines the wide, actively deforming margins o f the Adriatic region which roughly correspond to the mountain belts where previous tectonic activity was concentrated. Although thrust faults and crustal thickening dominate the geological structure of the circum-Adriatic orogenic belts, the present-day deformation does not follow this style. Active thrusting occurs offshore along the southern Yugoslavian and Albanian Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxwvutsrq zyxwvutsrqp Figure 4. Recent and active tectonic features of Italy, Yugoslavia, Albania and western Greece (from Philip 1983). Heavy lines indicate a fault active in the Plio-Quaternary. Finer lines indicate axes of folds active in the PlioQuaternary. Normal faults, such as those in central Italy, are indicated by hashing and major offshore lineaments are shown as dotted lines. zy zyxwvu zyx zy A c t i w tectonics o j the Adria fir region 943 coasts (e.g. Boore et al. 1981) and at the northern end of the Adriatic, but normal faulting dominates the active deformation of central Italy (Fig. 4). The occurrence of active extensional tectonics in parts of the Apennines, where the older structure is dominated by major thrust sheets responsible for the crustal thickening, has confused much o f the argument and literature dealing with the Adriatic. 3 Previous studies Early authors who examined the distribution of seismicity in the Mediterranean (Barazangi & Dorinan 190c): Papazachos i 973) recognized several relatively aseismic blocks. Whilst the 4 Data reduction Focal mechanisms for 51 earthquakes, including 45 new or revised first motion fault plane solutions determined by the authors, are presented in this study (Fig. 5). Solutions for earthquakes that occurred before 1963 are mainly from McKenzie (1972), and published centroid-moment tensor solutions for events occurring in 1983 and 1984 are also included. Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 study of seisniicity patterns is useful, reliable focal mectianisms are necessary to understand the motions between such blocks. The first such comprehensive study of the Mediterranean was made by McKenzie (1972) who constructed fault plane solutions for the largest events and publislied the polarities used in his mechanisms so that their reliability could be assessed. Ritsema (,I 975) summarized the present day deformation of the western Mediterranean but offered no explanation for the motions of areas with distinctive focal mechanisms (for example. thrusting in Yugoslavia relative to normal faulting in peninsular Italy). More recently, many poorly constrained or erroneous focal mechanisms have been published. These have usually been determined using polarity data published in agency bulletins (e.g. those o f the International Seismological Centre, ISC) without examination o f the relevant seismograms. Unless the polarity data can be independently checked, we do not consider such solutions to be reliable. and do not discuss them further in this study. Other studies in which mechanisms :ire presented without any polarity information are also not discussed unless such data are retrievable from other sources. The inversion of long-period body waves for the centroid-moment tensor (Dziewonski, Chou & Woodhouse 1981, section 4.3) has been widely used to obtain focal mechanisms. Giardini et al. (1984) published centroid moment tensors for 35 moderate to large earlhquakes that occurred in the Mediterranean since 1977. b u t did not attempt any kinematic or dynamic synthesis of the area from these data. Our study mainly updates the work of- McKenzie (1972, 1978). The additional fifteen years of data available since his 1972 study allows a much better appraisal of the seismicity in the western Mediterianean, and the use of seismic techniques developed since 1972, such as waveform modelling and relative relocation, provides a better understanding of the source geometry in large earthquakes and their aftershock sequences. Focal mechanisms of small earthquakes, especially aftershocks, often reflect minor internal deformation of the blocks bounded b y large seismogenic faults (e.g. Soufleris et al. 1982; Ouyed e f al. 1983; Deschamps & King 1984; King et al. 1985; Westaway & Jackson 1987). Inclusion of such mechanisms in a regional study only confuses the patterri of the larger scale deformation. For this reason, only earthquakes with body-wave magnitudes greater than about 5.2 have been considered in this study. Our objective is not to look at the details of each earthquake, but rather to see how the faulting in large earthquakes is related to the regional deformation. 944 zyxwvutsrq zyxwvutsrq zyxwv H. Anderson and J. Jackson Table 1 lists the events, as well as an index t o the figures showing their first motion polarities. The locations of these earthquakes are listed in Appendix 1 and details of their nodal plane orientations are included in Appendix 2 . For convenience, all earthquakes for which a fault-plane solution is shown are identified by number or date. as in Table 1. 4.1 LOCATIONS 4.1.1 Epicentres Most of the epicentral locations used here are those reported b y the International Seismological Centre (ISC) and by the National Earthquake Information Service ( N E E ) of the Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxwvu zy Figure 5. Fault-plane solutions for shallow earthquakes of the peri-Adriatic. Compressional quadrants are shaded and each event is numbered as in Table 1. P-axes are shown as a d o t in the dilatational quadrant and the horizontal projections of slip vectors are shown as arrows. Location and nodal plane information is given in Appendices 1 and 2. zyx zyxwvutsr zyxwvutsrqp zyxwvuts 94s Active tectonics of the Adriatic region Table 1. Earthquakes with mechanisms shown and discussed in this study. For hypocentral locations see Appendix 1 . ?.lo )lo F 16. P o l a r i t y Tcxt Area Source No. Date rime mh No source SI. Source I 2 i 4 5 6 8 !1 IC 11 S 5 5 5 11 I3 \'el. I1 5 . - slclly 5.6 5.6 5.4 5.2 I t a l y 5.4 Yugoslavia 5 . 6 W.(;reecc 5.4 5.5 - 13 13 5.S 5.5 - 14 5.5 Albania 5 . 1 SlCllY 5 . 1 SlCl I?. 5.1 5.6 5. 5 5.6 I\'. ( ; r c e c e 5.5 K.!;reccc 5.4 Yugoslavia 5.4 Yugoslavia 5.5 Albania 5.2 I t a l y 5.6 I< .!;reece 5 . 5 li.C;rccce 5.3 5.6 5.6 5 . 3 ri.Italy 5.3 K. I t a l y 5.3 K . I t a l y 5.3 & . I t a l y 5.1 S i c i l y 5.4 Yugoslavia 5.1 Yugoslavia 5 . 4 Yugaslav1a 5.4 Yugoslnvla 5.2 I t a l ) 5.1 S i c i l y 5 . 4 Yugoslavia 5.1 sic1l y 5.2 zyxwvutsrqpon zyxwvutsrq 620111 620821 630726 631236 0505 5 . - 1819 0417 1337 blfl4l.i O R 3 0 (16071)5 0201 1: 1;. 14 D-lIiO 680115 680116 680125 6S0528 15 16 20 21 22 23 24 25 2 (1 2; j.4 (ism 0102 1536 US10 0201 0133 1407 1552 - - - - 5.6 5.4 5. 1 5.3 5.6 5.5 5.3 5.2 5.2 5.6 5.8 '1 13 - - - - 1: 13 1: 13 15 15 13 11 1; 14 12 1 1: c C (. (: c c i' c (- i' c: C Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 690x18 691013 691026 69102; 700819 710715 ??!I917 ?311114 I !I 5.5 5.6 0-25 0 . 0 0201 5.1 1642 5.1 0956 5.1 Oy59 69040.3 2 2 1 2 17 18 - 1 1 1 zyxwvutsrqponmlkj 2s 29 30 31 32 i3 33 35 36 _,- 3 58 39 40 41 42 43 34 45 46 47 48 49 50 51 760911 -60911 760915 760915 :so415 7904(19 '90415 790415 790524 790919 791208 8nnsi8 800528 501123 810813 821116 83011830323 840429 840507 840511 840513 840709 1631 1655 0315 0921 2333 0210 0619 1443 1723 2135 0406 2002 1951 1834 11258 2341 1241 2351 0503 1749 1041 1245 1857 5.2 5.5 5.7 5.4 5.5 5.i 6.2 5.7 5.8 5.9 5.4 5.7 5.7 6.0 5.4 5.6 6.1 - - - ~ ~ - 12 3.11 19 6.04 2.24 17 18 I 1 I 6.92 17 1 12 12 12 7 13 1.3 5 13 11 1; 1 13 - 1.39 18 1 - ~ 2 - 8.85 3.84 2.43 3.90 3.20 2.35 2.23 3.4 7.8 2.0 1.7 7.6 c i' i' !' [ [ c: c c c c bI (: (: - (: c i: c (, S. 1 2 1 17 1 6 17 17 5 17 17 16 6 6 7 5 5 5 19 18 5.5 5.6 A1 banm li.Grccce 5.6 W .Greece 5.2 5.2 Italy Italy Italy Yugoslavia Greece c c zyxwvutsrqp 5.8 5.2 5.5 5.2 5.1 5.1 2 2 2 2 2 5.: 5.4 5.5 - __ Notes Date given as y e a r , month, day Tune g i v e n as hour, minute m ' d e r i v e d from USCS l i s t i n g s e i s m i c moment M' SF: s c a l e f a c t o r (10" ~ n ) Mo source: 1 G i a r d i n l e t a l . (1985) 6 Dziewonski e t a l . (1985) 7 l r b y e t a l . (1985b) ~~l F i g . i n d i c a t e s figure where p o l a r i t y observations a r e s h o m . i f no p o l a r i t i e s a r e p r e s e n t e d , figure refers to that showing only a shaded q u a d r a n t s f a u l t plane s o l u t i o n b u t p o l a r i t y s o u r c e colunm i n d i c a t e s r e f e r e n c e t o a v a i l a b l e p o l a r i t y informat ion. P o l a r i t y source: 1 blcKenzic (1972) 2 No p o l a r i t i e s , centroid-moment t e n s o r s o l u t i o n o n l y a v a i l a b l e . Rest: t h i s s t u d y . Tcxt: i n d i c a t e s s e c t i o n i n t h e t e x t where t h e event i s d i s c u s s e d . Source v e l . : C v e l o c i t y at f o c u s 6.8 ~ I / S ?I v e l o c i t y a t focus dependent on depth o f event hut o r i g i n a t e s i n t h e mantle. 946 H. Ariderson arid J. Jackson 4.1.2 Focal depths Errors in I'ocal depths determined from arrival times are usually greater than those in epicentral co-ordinates. Keliable focal depths can be determined ii' the surtace reflections p P or sP can be recognized in the seismogram: but these are rarely evident for earthquakes shallower than 7 0 km. The depths of shallow earthquakes can be obtained using either local seismograph networks o r synthetic waveform modelling. Local networks have been used successfully to determine aftershock depths in Italy and Algeria (Ouyed et al. 1983; Deschamps & King 1984) but such networks were generally not installed after the other large earthquakes considered in this study. In the last ten years the use of synthetic seismograms, pioneered b y Langston & Helmberger (1975). has become a standard technique for refining source mechanisms and focal depths. At moderate teleseismic distances and long periods, the early part of a seismogram generally consists of direct and surface-reflected phases. The relative amplitudes of these phases are determined by the focal mechanism, and their time separation is dependent o n the focal depth and velocity structure above the source. Other phases such as sea-bottom reflections and water multiples can cause additional waveform complexity. The most important parameters that affect the shape of the waveform are: ( I ) crustal velocity model, (2) focal mechanism, ( 3 ) source-time function, (4) depth. The crustal velocity models used in this study are mainly taken from local seismic surveys. Uncertainties in the orientation of the nodal planes in fault-plane solutions can be minimized by matching the relative amplitudes of the first two half-cycles of the waveform observed at stations of different azimuths. This method has been used t o check and refine the fault-plane solutions of those earthquakes whose seismograms we modelled t o obtain focal depth. The sourcetime function is taken ta b e a trapezoid with a rise, a plateau and a faall time. The time function is a major source of ambiguity for some shallow earthquakes because there is a trade-off between source-time function and focal depth (Kadinsky-Cade & Barazangi 1982; Christensen & Ruff 1985). However, the width of the first half-cycle in a waveform is Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxwv zyxwvutsr zyxw zyxw zyx zyxwv zyxwvu United States Geological Survey (here referred t o as USGS). Earthquake locations determined by the ISC are based on rnany more arrival times than those o f the USGS and are probably more accurate. Some reliable macroseismic locations are available and. where used, are discussed in the text. Earthquakes with reported magnitudes z 5.0 are usually recorded by rnany stations, and Ambraseys & Melville (1982) have shown that large events ( m b z 5.5) in Iran are generally located within 2 0 kin of their macroseismic epicentres. A similar accuracy is reported for earthquakes in Itaiy (Westaway & Jackson 1987). Algeria (Yielding ef al. 1981). and Greece (Soutleris & Stewart 1981 ;Jackson et al. lO82a). Relative relocation techniques can improve the accuracy of epicentral locations if the location of a reference event is known reliably, from either macroseismic evidence (damage distribution or recognition of a fault break) or a temporary local network. The relative relocation technique of Jackson & Fitch ( 1979) has buccesxtully refined the locations ( i f ' aftershocks i n Greece and Algeria (Jackson rt al. 19X?a; Yielding e f al. 1981) and is used here to improve the locations of a swarm of earthquakes that occurred in 1969 January in western Sicily. In this case, the location of the master shock was estimated from the maximum epicentral intensity distribution, which is quite localized for an earthquake of this relatively small magnitude (mh = 5 .I). This relocation method does not reliably determine focal depth because of the trade-off between depth and origin time [Jackson & Fitch l97Y). zyxwvuts zyxw zyxwvutsrq zyxwvutsr zyxwvutsrqpo zyxwvuts Active tectonics of the Adriatic region 947 generally most sensitive t o focal depth and the width of the first complete cycle is most sensitive t o the total duration of the time function. Synthetic waveforms for a range of depths are compared with the observed seismograms at various distances and azimuths and the best fit depth is adopted. The depths of large (mb > 5.0) earthquakes obtained by this method are probably accurate t o within 4 km (Jackson & McKenzie 1984). The scalar moment can be calculated by comparison of the observed and synthetic absolute amplitudes. 4.2 1; A U L T -P L A N E S 0 L U 'I 1 0 N S 4.2.1 First-motion solutions KEY: zyxwvuts zyxwvu zyxwvutsrqp 0 long period dilatation long period compression o short period or uncertain dilatation short period or uncertain compression nodal dilatation M Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 The first-motion fault-plane solutions presented in this study are based on polarities read from WWSSN seismograms. We have read all the polarities of new solutions ourselves, and critical or anomalous polarities presented by McKenzie ( I 972, 1978) or Jackson ( 1 979) have been checked. Station positions on the focal sphei-e of events that are assumed to have occurred within the crust were calculated using a P-wave velocity below the source of 6.8 km s-'. This approach differs from that of McKenzie (1972. 1978) who assumed a x + + 0 -6- nodal (polarity uncertain) P axis T axis slip vector horizontal projection of slip vector nodal compression Figure 6 . Fault-plane solution for the earthquake of 1972 September 17 (event no. 24), at 1 4 hr 07 min (GMT), t o show the symbols and conventions used in Figs 7 , 9 , 10, 1 1 , 12, 1 3 , 14, 15 and 17. Appendices 1 and 2 give details of location and timing of each earthquake, and specifications of the nodal planes. H. Anderson and J. Jackson mantle velocity below the source of most earthquakes in this region. We therefore replotted McKenzie’s fault plane solutions using focal spheres calculated with crustal source velocities. Thus the mechanisms shown here as McKenzie’s (1972) may be slightly different from those he presented even though they are based on the same polarity information. Although long-period vertical WWSSN seismograms were used for almost all polarity observations, some polarities were read on short-period vertical instruments when longperiod records were unavailable o r obscured. The short-period polarity observations were only used if the onset resembled the impulse response of the instrument. The long-period polarity observations were discarded if their arrival times were later than those observed on the short-period records. In cases where the polarities were critical to the nodal plane orientations, they were checked on the long-period horizontal components. It was frequently observed that the stations BUL and SHI had reversed instrumental polarities. All the fault-plane solutions in this study are equal area, lower focal-hemisphere projections, using symbols that are shown in Fig. 6, which serves as a key for the later figures. In cases where two or more mechanisms have been determined for the same earthquake, the solution with solid nodal planes is the one we prefer. Alternative solutions are shown with dashed nodal planes and are discussed in the text. The orthogonality of nodal planes that were determined graphically were checked using a computer program written by R. Westaway. Some of the solutions shown by McKenzie (1972, 1978) and Jackson (1979) have been adjusted accordingly. and the nodal planes listed here (Appendix 2) are therefore slightly different from those quoted in these earlier papers. Where only one nodal plane is well constrained (the second is usually shallow dipping in this case), the focal mechanism is shown as pure dip-slip. This is not intended t o imply that no strike-slip component is involved, but no better alternative can be provided without further evidence (from S-waves, waveform modelling, fault-break mapping, etc.). The constraints on the nodal planes are best assessed by examining the distribution of the polarity observations themselves. 948 zyxwvut Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxwvutsr zyxwvu zy 4.2.2 Centroid-moment tensor solutions Some ‘best double couple’ centroid-moment tensors (Dziewonski e t al. 1981) obtained from inversion of long-period body waves are available for some of the more recent events. Where both centroid-moment tensor and first motion solutions are available for the same earthquake, the difference between them is discussed. For recent (1983-1984) earthquakes only the centroid-moment tensor solutions are available. In general, the centroid-moment tensor solutions are in reasonable, though rarely perfect, agreement with first-motion polarities. In the absence of first-motion data, the centroid-moment tensor solutions must be interpreted with care, because of their inability t o resolve M,, and M y , components of the moment tensor for shallow events (see Scott & Kanamori 1985) and also because multiple events involving rupture on fault planes of differing orientation can lead t o an ‘overall’ moment tensor that cannot be directly compared with particular faults (see Berberian et al. 1984). 5 Fault plane solutions A large number of papers have been published on the focal mechanisms of several recent destructive earthquakes occurring in the Adriatic region (Skopje, 1963 July 26; Friuli, 1976 May 6; Montenegro, 1979 April 15; Campania-Basilicata, 1980 November 23) but n o study has been made of the relationship between these and other earthquakes, less signifi- zyxw zyx zyxwvuts 949 Active tectonics of the Adriatic region cant in human terms, for which focal mechanisms based on teleseismic first motions can be determined. Although there are several studies of the focal mechanisms of Italian earthquakes, these either use only polarity observations from local networks (for small earthquakes), or use polarities from ISC bulletins, which we d o not consider to be reliable. Fig. 5 shows the fault plane solutions that we consider reliable for large earthquakes in the Adriatic. Appendices 1 and 2 contain details of the locations and focal mechanisms. Lists of polarities are available fi-om the authors. 5.1 SICILY Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 The belt of seismicity and the main structural trends that occur along the north coast of Africa extend across the Strait of Sicily into the southern Tyrrhenian continental margin and Sicily itself. Fault-plane solutions for six earthquakes in Sicily are shown in Fig. 7. These events are shown in their regional context in Fig. 5. The most westerly event (no. 39, 1979 December 8) has a reasonably well-constrained focal mechanism, and if the N-S trending nodal plane is chosen as the fault plane then the slip vector shows general agreement with other solutions from this area. Further t o the east, three earthquakes (nos 13--15, 1968 January 15, 16, 25) are clustered in western Sicily. The earthquake swarm including these events was studied by de Panfilis & Marcelli (1 968) and Cosentino & Mulone ( 1985) who report maximum damage near the town of Gibellina and the Belice Rwer (Fig. 8), although no clear fault break was recognized. The relative locations of events nos 13-15 and five other large earthquakes that also occurred in 1968 January were determined using ISC arrival time data and the relative relocation technique mentioned earlier. The location of the reference shock ( 1 968 January 15; 0201 hr) was chosen as 37.75"N, 12.98"E. based on the maximum epicentral intensity on the isoseismal map for this event (Barbano et al. 1980). Fig. 8 shows the relocated positions of these events and Table 2 lists the new locations we determined. This swarm of events appears to be aligned N-S but cannot be related t o any recent faulting or major structural trend. The depths determined using this relocation technique cannot be considered reliable, but the pattern of epicentres is probably accurate t o about 5 km. The mechanisms for these events (nos 13-1 5) were determined by McKenzie (1972) who presented them as pure thrusts because the south-dipping plane was unconstrained. Some additional polarity observations have been made and several alternative solutions (dashed) are indicated in Fig. 7. In each case, polarity observations are satisfied by either a pure thrusting mechanism or one with a NNW striking plane that dips WSW and has a right lateral strike-slip component, The alignment of the relocated epicentres in a northerly direction suggests that the solutions with appreciable strike-slip component are preferable. If so, the slip vectors for events nos 13--15 are oriented approximately N-S. Two other large events occurred off the northern coast o f Sicily but within the area of continental shelf (no. 33. 1978 April 15; no. 41, 1980 May 28). The centroid-moment tensor solution for event no. 41 (dashed line in Fig. 7) indicates almost pure dip-slip thrusting but this solution is incompatible with several of the polarities observed t o the SE. Modelling of the teleseismic P waves for this event (Fig. 9) indicates a simple source at a depth of 12 km. The moment calculated from this modelling is 3.5 x IOl7Nm which is almost identical to that determined from the centroid-moment tensor inversion (3.84 x 1 017 Nm; Giardini et al. 1984). Event no. 33 (1978 April 15) was located just south of the island of Vulcano and has been studied by del Pezzo & Martini ( 1 982) and del P e u o er al. ( 1984). Del Pezzo & Martini zyx zyxwvu Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 W zyxwvutsrq H. Anderson and J. Jackson 950 zyxw zyx zy zy Active tectonics of the Adriatic region 95 1 38’ Figure 8. Relocated epicentres for the largest events in the 1968 January earthquake swarm in western Sicily. Open circles indicate ISC positions before relocation, and the number beside the filled circle identifies the event in Table 2. Thin h e s show structural lineaments recognized by de Panfilis & MarcelLi (1968). ( 1 982) relocated the aftershocks of this event, but the pattern of epicentres they obtained does not show a clear trend. Del Pezzo et al. (1984) relocated epicentres for the whole Aeolian region, using two different velocity models. There is a strong gradient in crustal thickening in this area with an increase in the depth of the Moho from about 2 0 km in the most northern part t o about 35 k m under Calabria and Sicily (Morelli et at. 1975). The first velocity model chosen b y del Pezzo et al. for relocation of the Aeolian earthquakes is characteristic of the northern part of the islands with a crustal thickness of 2 0 km. The second model places the Moho a t 28.7 km (‘a sort of average between two different structures’; del Pezzo et al. 1984). The two different velocity models lead t o very different hypocentre distributions, but neither of these relocations shows any alignment of the low-magnitude seismicity with any major structural feature. Table 2. Macroseismic location of mainshock (1968 January 15) and location of other earthquakes in W. Sicily relocated relative t o this event (see Fig. 8). MainshockNo. Lat. 37.807 37.676 37.830 37.817 37.750 37.793 37.857 37.687 Long. Depth(km) Mag. HI Date 13.012 12.966 12.983 13.006 12.983 12.960 12.976 12.966 19.0 1.0 22.0 34.0 13.0 23.0 36.0 3.0 5.1 5.0 4.7 5.1 5.4 4.6 5.1 5.1 1228.4 1315.7 1548.5 133.0 201.1 318.7 1642.7 956.8 140168 140168 140168 150168 150168 150168 160168 250168 zyx Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 37O zy 952 zyxwvutsrq zyxwv H. Anderson and J. Jackson 2.4 zy Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 3.2 zy zyxwvu zy zyxwvutsrqp II 'J 0I 60sec KBL 2.4 Figure 9. Observed (solid) and synthetic (dashed) waveforms for event no. 41 (1980 May 28) at a focal depth of 1 2 km. Synthetic waveforms were calculated using the solid nodal planes. WWSSN station code and moment (X 10" Nm) are indicated b y each waveform pair. A velocity model identical t o that used for m o d e l h g event no. 33 (Table 3) was used, and direct waves and reflections from each interface were included. A triangular time function of 1, 0, 1 s was used. zyxwvutsr The focal mechanism for event no. 33 (1978 April 15) was determined from first-motion data (solid and dashed lines) and the centroid-moment tensor inversion (dash and dotted line in Fig. 7; from Giardini et al. 1984). The nodal planes in the moment tensor solution violate two short-period compressional onsets near the null axis (Fig. 7). A third set of nodal planes, involving more strike slip motion but inconsistent with a nodal dilatational onset to the northwest, is shown by dashed lines in Fig. 7. To try and constrain this focal mechanism and determine the focal depth, we modelled the P-waves from this event (Fig. 10). The velocity model chosen was that determined by del Pezzo & Martini (1 982) with the Moho at 20 km. The nodal planes were adjusted t o fit the amplitudes of the first half cycle. The solution shown as the solid nodal planes in Fig. 7 is compatible with both these amplitudes and the polarity observations. It is clear, however, that this event did not involve a single. simple rupture. Second and third ruptures (or sub-events) with time delays of 6 and 15 s occurred, apparently t o the southeast of the mainshock (see Table 3 for details). There is no independent evidence for the orientation of faulting in these second and third sub-events so their mechanisms have been assumed t o be the same as that of the mainshock. The focal 780415 - Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 ii zyxwvu zyx zyxw zyxwvutsr zyxwvutsrqpon A A E 0.59 / I' L' 0 6Osec LAH 1.15 Figure 10. Observed (solid) and synthetic (dashed) waveforms for event no. 33 (1978 April 15) at a focal depth of 2 1 km. Synthetic waveforms were calculated for the solid nodal planes. WWSSN station code and calculated moment (X 101*Nm)are shown beside each waveform pair. Details of modelling parameters are shown in Table 3. 954 zyxwvutsrqp zyxwvutsrq zyxwvutsrqp zyxw H. Anderson and J . Jackson depth of all three sub-events is estimated at about 21 km. The seismic moment of 9.7 x l O I 7 Nm estimated from waveform modelling is in reasonable agreement with that of 1.39 x 1 0l8Nm obtained by inversion for the moment tensor by Giardini et al. (1 984). This event therefore involved complex thrust faulting which, because of the apparent rupture propagation towards the southeast, suggests that the NW-striking nodal plane may have been the fault plane. There is some independent evidence for NW-trending faulting in this area (Fig. 4 and the Tindari-Letojanni fault system of Ghisetti, Scarpa & Vezzani 1982). Slip vectors for each of the Sicilian earthquakes discussed here appear t o reflect the N-S directed convergence of Africa and Eurasia predicted in this area from the AfricaEurasia pole of rotation (see Fig. 2). zy zyxwvut 5.2 CENTRAL ITALY Table 3. Model parameters used for P-wave modelling of event no. 33 (1978 April 15). Subevent No. Strike Dip Rake Moment AT AX AY 148 148 148 55 55 55 154 154 154 1.0 1.0 0.8 0 6 15 0 -15 -17 0 7 15 1 2 3 AT is time delay, AX, A Y are north and east offsets in k m respectively. A triangular time function of 2, 0, 2 was used for all subevents. Velocity model Layer no. P-wave km s-' S-wave k m sel Density Mg m - 3 Thickness km 1 2 3 4 0.001 1.5 4.64 5.68 6.58 7.85 0.001 0.001 2.67 3.27 3.8 4.53 0.001 1.0 2.5 2.8 2.9 3.3 0 0.7 4.3 5 .0 10.0 25.0 5 6 Stations QUE KB L AAE KEV KBS LAH SHL Distance (degrees) Azimuth Ray paramete1 43.17 43.20 36.17 32.08 40.54 48.33 64.70 84.76 77.81 136.73 7.79 359.08 77.87 76.58 0.072 0.071 0.075 0.077 0.073 0.069 0.057 Direct waves and reflections from each interface were modelled. Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 Fault-plane solutions for seven earthquakes on the Italian peninsula are shown in Fig. 5 . Of these. three (nos 47-49) occurred in 1984 and insufficient seisnlograms were available f o r firstmotion fault-plane solutions t o be determined. The most southerly events, in the zyxw zyx z zyxwvutsrqp zyxwv zyxw Active tectonics of the Adriatic region 21-AUG-62 18:ltl 15-JUC-71 81 :33 23-NOV-80 18:34 Figure 11. Fault plane solutions for peninsular Italy. The solution for event no. 5 (1962 August 21) is from Westaway (1987) and the polarities for no. 42 (1980 November 23) are from Westaway & Jackson (1987). Symbolsas in F i g . 6. Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 19-SEP-79 21 :35 955 zyxwvu Campania area, have recently been studied in detail by Westaway (1987) and Westaway & Jackson (1987). Polarities for four of the Italian events (no. 42, 1980 November 23; no. 23, 1971 July 15; no. 38, 1979 September 19;no. 5 , 1962 August 21) are shown in Fig. 11. The Campania-Basilicata earthquake (no. 42) was the largest (M,= 6.9) to have occurred in peninsular Italy this century. The mainshock mechanism and aftershock distribution have been studied by Westaway & Jackson (1987). Deschamps & King (1983, 1984) and many others (see Westaway & Jackson 1987 for a review). Surface faulting in this earthquake shows that the NE-dipping nodal plane was the fault plane (Westaway & Jackson 1984). Aftershock studies by Deschamps & King (1983 and 1984) support this choice of fault plane. P and SH waveform modelling shows that this event was a complex multiple rupture nucleating at about 10 km depth (Westaway &Jackson 1987). The fault-plane solution of another earthquake which occurred in the Campania region (no. 5 , 1962 August 21) was published by McKenzie (1 972). His first-motion polarities have 956 zyxwv zyx zyxwvutsr zyxwvu H. Anderson and J. Jackson Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 been replotted with a crustal velocity at the focus and a solution very similar t o that determined by Westaway (1987; fig. 11) is shown in Fig. 5 . Relocation of this event, along with local geological structure seen in seismic reflection profiles, suggest that the NE-dipping nodal plane is the fault plane, and the waveform modelling indicates that both this earthquake and its large aftershock nucleated at about 8 km depth (Westaway 1987). This modelling and the depth of 1 0 km determined for event no. 42 (1980.1 1.23) indicate that extension in the southern Apennines occurs by steep normal faulting in the upper 10-1 5 km o f the crust. In 1984 May, two earthquakes occurred in central Italy in the Abruzzi-Lazio area (no. 48, 1984 May 7 ; no. 49, 1984 May 11). The first of these was relatively large (M, = 5.8) and caused extensive damage in the Abruzzi area. The centroid moment-tensor solutions for both these events (Dziewonski et al. 1985) show normal faulting with an orientation very similar to that in event no. 5 (1962.8.21). These events are located near some major N-S and NW-SE trending faults (Fig. 4), but there is no conclusive field evidence t o favour either nodal plane as the fault plane. This ambiguity leads to an uncertainty in the slip vector direciion of event no. 48 of about 50". Moments of 7 . 8 2 ~ l O ' ~ N and m 2 . 0 3 ~1017Nm were determined by Dziewonski et al. (1985) for events nos 48 and 49, respectively. The Norcia, or Umbrian, earthquake 1979 September 19 (no. 38) occurred very close to the Anzio-Ancona line (Fig. 4) which trends approximately NNE and separates the northern and southern Apennines. The first-motion fault-plane solution for this event is shown in Fig. 11 along with the moment-tensor solution of Giardini et a f . (1984), which is shown by dashed nodal planes. First motion solutions have also been determined by Deschamps, Iannaccone & Scarpa ( 1 984), Gasparini etal. ( 1 980) and Gasparini. Iannaccone & Scarpa (1985). All of these solutions show dominantly normal faulting. The first-motion solution here is well constrained, and inconsistencies of first motions with the centroid moment-tensor solution may indicate source complexity. Deschamps et el. (1984) located the aftershocks of the Norcia earthquakes. They found a t least two clusters of aftershock activity, which they thought revealed an elongated pattern of seismicity extending for about 8 k m parallel t o the Apennine trend. They then used this aftershock distribution t o suggest that the nodal plane striking NW in the fault-plane solution of the mainshock was the fault plane. However, the trend in the aftershock distribution is weak and we believe that the choice of fault plane is unresolved. Although normal faulting farther south in the Campania region occurs on NW striking faults, the proximity of the Norcia earthquake t o such a major structural feature as the Anzio-Ancona Line may suggest movement on a NE striking fault. Lavecchia, Minelli & Pialli (1984) report recent left lateral motion on NNE-SSW shear zones in the Umbrian area. If the NE-striking nodal plane in the fault-plane solution was the fault plane, then it too would have a component of left lateral strike-slip motion. The slip vector on the NE-striking plane is similar in direction to that in the large normal-faulting events further south. A seismic moment of 6.92 x 1 OI7 Nm was determined by Giardini et al. ( 1 984), which compares well with the value of 7.0 x l o L 7Nm determined by Deschamps et al. (1984). The earthquake of 1984 April 29 (no. 47), near the Tiber River, occurred in an area of NW-SE trending, recently active, grabens (Fig. 4). Only a centroid moment-tensor is available for this event (Dziewonski et al. 1985). It shows normal faulting with only a small component of strike-slip, so the slip vector direction is almost the same on both nodal planes. It is probable that the steeper nodal plane was the fault plane, as in the Campania area further south (Westaway 1987; Westaway & Jackson 1987). The earthquake of 1971 July 15 (no. 23) occurred further north near the Po River. Its fault plane solution (Fig. I 1 ) isnot particularly well constrained. although a similar solution is zyxwvutsr zyxwvuts zyxwvutsr zyx zyxwvu zyxwvutsrq zy zyxw Active tectonics of the Adriatic region 957 also reported by Gasparmi et al. ( 1 985). Major structural trends change from a NW-SE t o an E-W trend in this area but there is no clear reason for choosing either nodal plane as the fault plane. The slip vector on the E-W nodal plane is very similar t o those of the other large earthquakes in the Italian peninsula. 5.3 N O R T H E R N ITALY 5.4 YUGOSLAVIA Recent large earthquakes have been concentrated in three main zones in Yugoslavia: the Banija area of northern central Yugoslavia (near 45"N, 17"E); the Dalmation coast south of Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 North of the Po Valley, epicentres are not concentrated in clearly defined zones (Fig. 3), but in the last twenty years the seismicity pattern has been dominated by a swarm of earthquakes in the Friuli area (Fig. 4). Five events large enough for fzult-plane solutions to be determined occurred within four months of the largest shock of this sequence (1976 May 6. no. 26). The Friuli earthquake swarm has been the subject o f many studies (e.g. Cagnetti & Pasquale 1979; special issue of Bollettino di Geofisica, vol. XIX, 72, 1976). and focal mechanisms have been presented by Cipar (1980, 1981). We collected some additional polarity observations and our solutions for these events (no. 26, 1976 May 6: nos 29-32, 1976 September 1 l a , b and 15a. b) are presented in Fig. 12. Our polarity observations for the mainshock (no. 26) are shown with the nodal planes determined by Cipar (1980). because the shallow dipping plane in his solution is constrained by SH and SVwaveforms. Cipar (1980) calculated a focal depth of 8 * 2 km and a seismic moment of 2.9 x Nm for the mainshock from long-period P-wave modelling. A similar depth of 6.5 kin was estimated by Zonno & Kind ( 1 984). using depth phases identified at regional distances by the Grafenberg array. Four aftershocks of the Friuli earthquake (nos 29 -32) were large enough for first-motion solutions t o be determined. Event no. 29 is not well constrained, but observed polarities are consistent with a mechanism almost identical with that of the mainshock. The steeply dipping plane of event no. 30 is well constrained but the choice of a shallow dipping plane as the fault plane is again arbitrary. This mechanism is shown as a pure thrust. in spite of a dilatational onset at TRI. because the close proximity of this station t o the epicentre (0.76") makes its position on the focal sphere highly dependent on local velocity structure. In contrast t o the three large earlier events, first motion polarities for both the two aftershocks occurring on 1976 September 15 (nos 31, 32) require a component of strike-slip motion. Both of these mechanisms are well constrained, and Cipar (1980) determined a depth of 6 km for event no. 32. No surface faulting was observed following any of these earthquakes, and so their correlation with specific faults has been difficult. Weber & Courtot (1978) recognized several trends of faulting, but with E-W striking thrusts the most dominant. They suggested that movement in the mainshock (no. 26) occurred on a NNE striking, left-lateral strike-slip fault. The S-wave data of Cipar ( 1 980). which control the orientation of the shallow dipping nodal plane for this event, suggest that this interpretation is incorrect, arid that faulting probably occurred on one of the E-W striking thrusts dipping shallowly to the north (Fig. 4), which have been recognized in this area. The northward dipping nodal planes for aftershock nos 3 1 and 32 are steeper than in the earlier three events, but the slip vectors on these nodal planes are all similar and are oriented approximately NNW. This slip-vector direction is in complete contrast t o that observed in peninsular Italy. 958 zyxwvutsrq zyxwvuts zyxwvu Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 Figure 12. Fault plane solutions for the Friuli earthquakes. Symbols as in Fig. 6 . zyxwvutsrqp H. Anderson and J. Jackson Active tectonics of the Adriatic region 9-APR-79 zyx zy 959 zyxwvutsrqpon zyxwvutsrqpon zyxwvu 8 2 : I0 15-APR-79 & I 9 + 3 . b 0 I Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 I I * 0 . zyx zyx Figure 13. Fault plane solutions for earthquakes in Yugoslavia. Symbols as in Fig. 6 . Split; and an area of southern Serbia and Macedonia near Skopje (Figs 4 and 5). First-motion poiarities for some of these events are shown in Fig. 13. 5.4.1 Northern central Yugoslavia The most northerly cluster of seismicity occurs south of Zagreb near the Save River, where a set of active faults crosscuts the NW-SE structural trend of the Dinarides (fig. 4, Cvijanovic & Prelogovic 1977). There is a marked difference between the strike-slip mechanisms of three closely grouped events, no, 20 (1969 October 26). 21 (1969 October 27), and 43 (1981 August 13), and the more northeasterly thrusting mechanism of event no. 8 (1964 April 13). The first-motion solution of event no. 2 0 is not well constrained and polarities could also be consistent with an E-W striking thrust. However, the similarity of the polarities observed for this event and for the better-constrained event no. 21, which occurred less than a day later, suggest that a strike-slip solution is more likely. Similarly, event no. 43 960 zyxwvutsr zyxw H. Anderson and J. Jackson is not well constrained by the first-motion polarities, but a strike-slip solution is very similar to the centroid-moment tensor solution determined by Giardini et al. (1984). The mechanism for event no. 8 could not, however, be dominated by strike-slip motion. MdKenzie's (1972) solution for this event (the dashed nodal planes in Fig. 13) has been redrawn t o eliminate several inconsistent polarities. The resultant almost pure thrust reflects a different style of deformation from the strike-slip motion occurring less than 100 km t o the southwest. The NE slip vector direction for event no. 8 is similar t o that in the strike-slip events, if the NE-SW striking nodal planes are chosen as the fault planes. Young structural features support this choice (Fig. 4) and it seems likely that the NE-SW trending faults in this area are now active with left lateral strike-slip motion. We are not aware of any reports of surface faulting associated with specific earthquakes in this region. 5.4.2 Serbia and Macedonia 5.4.3 Dulrnutiuti coast The southern Dalmatian coast is the most seismically active area of Yugoslavia (Fig. 3). Earthquakes occur in a band approximately 100 km wide running south from near Split towards Albania. The historical seismicity of this area shows a similar pattern with the largest events occurring south of Split (Anderson 1985). The recent seismicity is dominated by the 1979 April 15 Montenegro event and its aftershock sequence, but two important events have occurred further north along the coast. The fault plane solution for event no. 4 (1962 January 11) was presented by McKenzie (1972), based on short-period observations from Di Filippo & Peronaci (1962). This solution is reasonably well constrained and is very Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyx zyxwv Three other major events have occurred in inland Yugoslavia in the last 22 years. Event no. 6 (1963 July 26) occurred very close t o the city of Skopje in Macedonia and has been the subject of many studies (e.g. Unesco earthquake study missions report 1968; Berg 1964; Arsovski et al. 1966; Ambraseys 1966). A first-motion solution for this event was determined by McKenzie (1 972) and has been redrawn in Fig. 13 using a crustal velocity at the focus. No surface faulting was observed after this earthquake (Ambraseys & Morgenstern 1966) but Zatopek (1968) suggested that the aftershocks align along a NW-SE trend in the vicinity of Skopje. However, this aftershock distribution is far from clear, and Sorsky (1 968). Arsovski & Hadzievsky (1 970) and Arsovski ( 1 970) report a clearly defined young zone of 'intense and sharply differentiated movements' running approximately ENE south of Skopjc. The strike of the SE-dipping nodal plane in Fig. 13 is not well constrained and a solution with a more ENE striking nodal plane could be constructed. We think it likely that this earthquake involved right-lateral strike-slip motion on a SE-dipping plane with the slip vector direction approximately NE-SW. This strike is similar to that of the Scutari-Pec Line (Fig. 4) and may indicate that this feature persists as an important tectonic boundary. Further north, a strike-slip mechanism was also determined for event no. 40 (1980 May 18). The first-motion solution is not well constrained and could also be drawn as a thrust (dashed and dotted line in Fig. 13). However, our preferred choice of a strike-slip solution is supported by the centroid-moment tensor solution of Giardini et al. (1984; dashed line in Fig. 13) and by another nearby strike-slip centroid-moment tensor solution (1984 September 7). which has nodal planes striking 278" and 10" and dipping 78" and 81", respectively (Irby et al. i985a). The main structural features in the epicentral region of these events follow the Dinaride trend (NW-SE), but we d o not know whether the NW or N E striking nodal planes are the fault planes. zyxwvu zyx zyxwvuts zyxwvutsr zyxwvut Active tectonics o j t h e Adriatic region 961 5.5 ALBANIA AND NORTHWEST GREECE Both normal and thrust faulting mechanisms occur in Albania and northwestern Greece. The thrust faulting mechanisms are concentrated along coastal Albania whereas the normal Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 different from the almost pure dip-slip events observed farther south (Fig. 5). One of the nodal planes strikes parallel to the coast and it is tempting t o assume that this one is also the fault plane. However. this choice involves a slip vector orthogonal to those observed elsewhere, and movement on the NNE striking plane is perhaps more likely. The centroid-moment tensor solution for earthquake no. 50 (1984 May 13), south of event no. 4. shows almost pure thrusting with nodal planes striking parallel to the coast and a seismic moment of 1.69 x 1 O i 7 Nm (Dziewonski e f al. 1985, and Fig. 5). The mechanism is very similar t o those of the Montenegro earthquakes further south. There are several studies of the Montenegro earthquake of I979 April 15 (no. 35), its foreshocks (including no. 34) and aftershocks (including nos 36 and 37). (e.g. Academy of Sciences, Albania 1983; Hurtig & Neunhofer 1980; Console & Favali 1981). A large foreshock (no. 34, m h = 5.3) occurred six days before the main shock. The first-motion solution for this event is very poorly constrained but it is important because i t requires a different mechanism from the following large events (nos 35-37). A solution can be drawn (Fig. 13 dashed line) so that one of the slip vectors is parallel to those observed for the following events, but this suggests movement on a N-S striking plane for which there is n o evidence from either the local tectonics or other fault plane solutions. A nodal plane with a shallow NE dip, similar t o that observed in the later events can be drawn, but the slip vector on such a plane is quite different from those determined for the other events. An active fault with a NE trend has been recognised in the vicinity of the Scutari-Pec Line further south (Kociaj & Sulstarova 1980) and this foreshock might have involved motion on a fault of this strike (solid line, Fig. 13). Several methods have been used t o determine a focal mechanism for the 1979 Montenegro mainshock (event no. 35). Our first-motion solution is shown as a solid line in Fig. 13. The mechanism determined by Boore et al. (1981) was based on P-waves and S-wave polarization (dashed line), and the centroid-moment tensor solution of Giardini et al. (1984) is shown by dashed and dotted lines. A slight refinement t o the solution was made using long-period surface waves (Kanamori & Given 1981) but is not presented here. The moment determined by the centroid-moment tensor method was 3.1 1 x lo'' Nm. A focal depth of 22 kin was estimated from modelling of long-period body waves by Boore et al. (1981). who found that the seismic moment estimated from the amplitude of the first cycle of long-period body waves was four times smaller than that calculated from inversion of the Rayleigh waves (4.6 x lo'' Nm). This discrepancy suggests that the event was a multiple rupture. A centroid-moment tensor solution (Giardini et al. 1984) for a large aftershock on the same day as the mainshock (1979 April 15, no. 36) with a seismic moment of 6.04 x 1017Nm is shown in Fig. 5. The first motions of this aftershock were obscured and unreliable. Another large aftershock occurred more than a month later ( I 979 May 24, no. 37). Two solutions that are compatible with the first motion polarities are shown in Fig. I3 (solid, dashed and dotted lines). In both cases the shallow dipping nodal plane is very poorly constrained. The centroid-moment tensor solution of Giardini et aZ. (1 984) is shown as a dashed line in Fig. 13 and, although there is notable inconsistency with some first motion polarities the strikes of the nodal planes are very similar. The slip vector direction is again approximately NE-SW and the moment determined for this event was 2.24 x 10"Nm. 962 zyxwvutsrqp zyxwvutsr zyxwvu zyxw zy H. Anderson and J. Jackson faulting events described here are at the westernmost limit of the extension in the Aegean area (McKenzie 1978; Jackson, King & Vita-Finzi 1982b). 5.5.1 Normal faulting Four dominantly normal faulting events are shown in Fig. 5. McKenzie's (1972) mechanism for the most northern of these (no. 12; 1967 November 30) has been redrawn with a crustal velocity at the focus and additional polarity observations (Fig. 14). The WNW dipping nodal plane is well constrained but the strike of the other is not. Surface faulting in this earthquake was described by Sulstarova & Kociaj (1980). The faulting had a strike of 040" and was more than 10 km in length. However, there is some confusion over the direction of throw o n this fault. Although Sulstarova & Kociaj (1980) emphasise that the strike of the surface faulting coincides with one of the nodal planes in their (apparently upper hemisphere) focal mechanism, photographs and text are ambiguous (e.g. 'All along its length Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 19-AUG-70 82 al zyxwvutsrqpo 4-NOV-73 IS:% zy Figure 14. Fault plane solutions for earthquakes in northwest Greece and Albania. Symbols as in Fig. 6. zyxw zyxwvu Active tectonics of the Adriatic region 963 the northwestern block dipped against the southeastern block'). Two solutions are presented here. The preferred mechanism (solid lines) does not violate the observed compression in the eastern quadrant, but the dashcd solution allows a SE-dipping plane with a similar strike t o the surface faulting observed by Sulstarova & Kociaj (1 980). It is not possible to identify the fault plane until the ambiguity in the field report is resolved. An almost purely dip-slip normal mechanism was determined for event no. 5 1 (1984 July 9; Fig. 5) using the centroid-moment tensor method (11-by et al. 1985b). This earthquake was sniall (seismic moment 7.6 x I O l h N m ) and no f'urthet- intotination is yzt available t o relate it to the regional tectonics. Two more normal faulting mechanisms have been determined for events further south (nos 9, 11 1966 February 5, 1967 May 1. respectively). Event no. 9 occurred very close to the Kremasta Dam in western Greece and is thought to be associated with the filling of the reservoir. The location used here was from macroseismic data (Soufleris 1980). A focal mechanism similar t o that presented here (Fig. 14) was also deterniirted by Stein, Wiens & Fujita (1982, using body- and surface-wave data), McKenzie ( I 972), and Fitch & Muirhead (1974). The depth of this event was estimated at 5 km from waveform modelling (Fig. 15a) and the seismic moment was about 5 . 0 ~ IOl7Nm. The depth of' S kin estimated here is much less than that of 15 kin estimated by Stein et al. (1Y81) but these authors d o not present sufficient waveform data for comparison with those shown in Fig. 15(a). It is interesting t o note that the Koyna earthquake of 1967 Decernber 12 in India, which is also thought to be related to reservoir filling, had a shallow focal depth of 4.5 km (Langston & Francn-Spera 1985). Both Stein c i a / . (1982) and Fitch & Muirhead (1974) suggested that the southerly dipping nodal plane was the fault plane, but this conclusion was based on either ISC o r relocated aftershock depths, which are unlikely t o be reliable. McKenzie's (1972) fault plane solution for another, more northerly, normal faulting event (no. 11, 1967 May 1) has been redrawn with a crustal source velocity (Fig. 14), and placed in its macroseismic location (Soufleris 1980), which lies west of the highest Pindos mountains. No surface faulting has been reported for this event, and the approximate N-S regional tectonic trends offer n o preference for choice of fault plane. Waveform modelling suggests that this earthquake nucleated at a depth of 11 km and had a moment of 1.25 x IO"Nm(Fig. 15b). This belt of normal faulting can only be placed in its regional context through comparison with other normal faulting earthquakes farther east in the Aegean. McKenzie (1978) and Jackson et al. (1982b) buth discuss the Aegean seismicity and no further synthesis has been attempted in this study. 5.5.2 Thrust faultiril: The belt of thrust faulting along the southern Dahiatian coast continues south along the coastal regions of Albania and northwest Greece. This area has been the site of intense historic seismic activity in the last 2000 yr. This coastal seismicity changes character south of the island of Kefallinia (Figs 4 and 5) and earthquake activity in that region is closely related t o the subduction in the Hellenic 'Trench. Focal mechanisms of six events from the coastal regions of Albania and western Greece are shown in Figs 5 and 14. The strike of the shallow dipping plane for the most northerly of these (no. 22, 1970 August 19) is not well constrained and is drawn here as a pure thrust. The main structural features in the vicinity of' this epicentre trend approximately NW-SE, parallel to the nodal-plane strikes shown here. It was not possible t o determine a first-motion solution for event no. 44 (1982 November Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxw zy zyxwvu zyxwvut 964 zyxwvutsr zyxwvutsrq H. Anderson and J. Jackson Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxwvutsrq zyxwv zyxwvut Figure 15. (a) Synthetic and observed (dotted) waveforms for 1966 February 5 for a depth of 5 km computed using a time function of 2, 0, 2 s and a velocity of 6.1 km s-' above the source. WWSSN station codes and the estimated moment ( X 10'' Nm) are shown next to each waveform pair. The average seismic moment is 5.0 X 10" Nm. (b) Synthetic (top) and observed (bottom) long period waveforms for 1967 May 1. Synthetics are calculated at a focal depth of 11 km and moments at each station are in units of 10" Nm. A time function of 2 , 0 , 2 s was used. zyx zyxw Active tectonics of the Adriatic region 965 Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 16) because almost all long-period P-wave onsets were confused by noise. However, a centroid-moment tensor solution has been determined by Giardini et al. (1984) and is shown in Fig. 5. The seismic moment for this event was 3.2 x 1O”Nm. The mechanism shows predominantly thrust faulting with a strike compatible with regional trends. Event no. 17 (1969 April 3) has been included because Sulstarova (1980) and Aliaj (1982) reported surface faulting related to this event. Aliaj (1982) suggests that movement took place on a NNW trending fault and Sulstarova, Kociaj & Aliaj (1982) report that they ‘measured the length of surface faults and dimensions of the pleistoseismal zones’ of earthquakes including event no. 17, although they do not present any details. Sulstarova (1982) gives a strlke-slip solution for this event (shown as a dashed line in Fig. 14). Only a limited number of long-period first-motion polarities were large enough to be read with confidence, and these are shown in Fig. 1 4 along with nodal planes indicating pure thrusting, which is compatible with these polarities and with the observed fault trend striking NNW mentioned by the Albanians. Several of these polarities are incompatible with the solution of Sulstarova (1982), so it seems likely that this event, like others in this part of Albania, involved thrust faulting on a NW-striking fault plane. A fault plane solution for event no. 19 (1969 October 13) was published by McKenzie (1972), but the wrong epicentral location was used for calculating the position of stations on the focal sphere. The solution in Fig. 14 has been recalculated at the macroseismic epicentre (Soufleris 1980) and uses additional polarity observations. The strike of the shallow-dipping plane is not well constrained but can be constructed to give a slip vector on the steep plane that is similar to those in nearby thrusting events. Although E-W, apparently left lateral, strike-slip faults of Tertiary age are known in this part of western Epirus (I.F.P. 1966), they are not well dated and there is no direct evidence of their recent reactivation, even by microearthquakes (King et al. 1983). Another, relatively poorly controlled, focal mechanism has been determined for an event (no. 25, 1973 November 4), w h c h occurred offshore at the junction of the Albanian thrust belt and the Hellenic Trench (Figs 5 and 16). Two alternative solutions are shown in Fig. 14. The predominantly thrust faulting solution (also shown by McKenzie 1978) is similar to those further north, but a large component of strike-slip could also be involved. However, if movement occurred on a NE-striking fault then this would require left-lateral strike-slip, as opposed to the right-lateral movement on NE striking faults shown by better constrained solutions further south (nos 24 and 26). We therefore think that the thrusting solution is more likely. Ambraseys (1975) observed surface faulting of an ambiguous nature that might be associated with event no. 10 (1966 October 29), and although the shallow-dipping plane is unconstrained in the focal mechanism (extremes shown as dashed or dot-and-dashed lines in Fig. 14), it can be constructed with a ”W strike, parallel to the surface faulting, without violating any of the observed polarities. The fault observed by Ambraseys had a length of between 2 and 4 km and a maximum vertical displacement of 0.4 m (Fig. 16). If this really was the causative fault, the slip vector determined from the fault plane solution (Fig. 14) trends 1134 which is unlike other thrusting events further north along the Albanian coast. Both this event and no. 19, occurred in a zone between the inland normal faulting and the coastal belt of thrusting, and where complicated faulting might perhaps be expected. The additional focal mechanisms presented here support the earlier assertion of McKenzie (1978) that the boundary separating the normal and thrust faulting in this area is apparently very sharp. McKenzie (1978) suggested that the complete change in stress orientation could be explained by the detachment and sinking of the lower part of the lithosphere beneath the thickened thrust zone. This model suggests a very limited area of compression along the 966 zyxwvutsr zyxw zyxwvu H. Anderson and J. Jackson Albanian coast, but the zone of thrusting clearly continues much further north along the Yugoslavian coast (where there is n o active inland belt of extension). 5.6 NORTHWEST HELLENIC SUBDUCTION ZONE Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 The region of the Ionian islands (Fig. 16) marks the northwest termination of the Hellenic subduction zone. Although there is no clear bathymetric expression of the subduction zone at its NW extremity (near the island of Kefallinia) the locations of large earthquakes and their mechanisms help in defining the actively deforming zone, The locations of the events described in this section are shown in Fig. I6 and the mechanisms for which polarity information is available are shown in Fig. 17. The two most northerly earthquakes in this zone occurred just to the west of Kefallinia and their epicentres lie above the steep-sided NNE-SSW trending bathymetric feature; the Kefallinia Valley (Fig. 16). The focal mechanism for no. 24 (1972 September 17) was shown by McKenzie (1978) as a pure dip-slip thrust. This is incompatible with several additional zyxwvutsrq Figure 16. Bathymetric map of the NW end of the Hellenic Trench, showing the epicentres of recent large earthquakes (numbered as in Table 1). The strike of surface faulting found by Ambraseys (1975) and possibly related t o event 10 is shown by a line through its epicentre. Bathymetry is from the International Bathymetric chart of the Mediterranean, UNESCO, 1981. zy zyx zyxwvutsrqp zyxwvutsrq zyxwv Active tectonics o f the Adriatic region 28 nAR 68 07 17-SEP-72 I 3 : 0 7 12.41 8 JUL 69 88 89 zyxwvutsrqpo zyxwvutsrqponmlkjihg I 1 -nA1-76 17:w 12 J’JN 76 0’2 59 23-flAR-83 23:51 Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 JAW83 967 zyxwvutsrq Figure 17. Fault plane solutions for earthquakes associated with the Hellenic Trench in western Greece. Symbols as in Fig. 6 . first motions that have become available, which require a well constrained strike-slip solution (Fig. 17). The NE-trending nodal plane is approximately parallel to the trend of the Kefallinia Valley, and we suspect that this mechanism involved right-lateral strike-slip in this direction. The centroid-moment tensor solution for the nearby event no. 46 (I983 March 23) is very similar, and compatible with the few polarity observations so far available (Fig. 17). The seismic moment determined by Dziewonski, Friedman & Woodhouse (1983) was 2.23 x 10’’ Nm. This mechanism also suggests motion on a right-lateral strike-slip fault that probably terminates the Hellenic subduction zone. The epicentre of event no. 45 (1983 January 17) also appears to lie on the southern extension of the steep bathymetric slope marking the eastern edge o f t h e Kefallinia Valley. However, a strike-slip mechanism cannot easily be drawn for this earthquake. The strike of the shallow dipping nodal plane is not well controlled and the mechanism here (Fig. 16) is drawn as a pure dip-slip thrust. A centroid-moment tensor solution has also been determined for this event (dashed line in Fig. 16) and although there are some discrepancies with the 968 H. Anderson and J. Jackson observed first-motion polarities, the solution has very little strike-slip component. This is the most northern event whose mechanism shows shallow angle thrusting of the Ionian Sea beneath Greece. Similar thrusting mechanisms for the nearby events nos 3 and 3 (1953 August 12 and 1959 November 15) were presented by McKenzie (1972) and are shown in Fig. 5, but their locations could be in error by as much as 5 0 km and either or both could reflect movement on thrust faults such as those observed on Kefallinia or Zakynthos (Mercier et al. 1976, 1979). One recent event appears to have occurred east of Zakynthos (no. 16, 1968 March 28). Its mechanism is not well constrained but polarity observations allow nodal planes striking NW, parallel to the fault strike observed on Kefallinia by Mercier et al. (1976) and in the offshore eastern basin (Brooks & Ferentinos 1984). Three events close together south of Zakynthos (nos 18, 2 7 , 2 8 ; 1969 July 8, 1976 May 11, 1976 June 12) also have dominantly thrusting fault-plane solutions (Fig. 17). In all of these solutions the shallow dipping nodal plane is unconstrained and so all these mechanisms are shown here as pure dip-slip events. An alternative solution is shown for event no. 27, but it is likely that all these mechanisms involved thrusting of the Ionian Sea beneath Greece. The most southerly event included in t h s study (no. 7 , 1963 December 12), is problematic. The three dilatational first motions in the SE quadrant (Fig. 17) preclude a mechanism similar to the others described here unless the northerly dipping nodal plane has a more E-W strike than shown. Two solutions with maximum strike slip components for this event are shown in Fig. 17 and, since there is no evidence for either, a pure dip-slip thrust is shown. This mechanism is not easily compatible with any regional interpretation but, because it is clearly different from the other thrusting events to the north, may indicate motion on a transverse structure in the Hellenic Trench. zyxwvutsrq zyxwvu Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 5.7 zyxwv zyxwvutsr MESSINA E A R T H Q U A K E (1908 DECEMBER 28) The 1908 December 28 Messina earthquake (event no. 1) is important because it is the only large recent event in the Calabria-Sicily region of southern Italy for which we have evidence of a normal faulting mechanism. In spite of its occurrence at a time when there were few seismological observatories operating, at least two focal mechanisms have been determined for this event (Fig. 5 , Gasparini et al. 1982; Schick 1977). Neither is particularly well constrained but both require dominantly normal faulting. This event is clearly unrelated to the Tyrrhenian Benioff Zone since thrust faulting would be expected if subduction were still occurring; it has a completely different mechanism from other events in Sicily and is quite different in orientation from other normal faulting events in peninsular Italy. The earthquake was large (magnitude 7 , Schick 1977) and produced a sizeable tsunami (maximum 10.6 m, Ryan & Heezen 1965) but no surface faulting. However, a spirit levelling survey had been completed a few months before this earthquake and was repeated immediately after it. These measurements were interpreted by Mulargia & Boschi (1982) to show a strong net subsidence of the Messina Strait area. From analysis of the same geodetic data, Schick (1977) suggested that the earthquake occurred on a NE-SW trending fmlt running through the Messina Strait. Recent tectonic features in the Calabrian area have a similar trend (Fig. 4) and are quite different in orientation from those faults in the rest of peninsular Italy. This earthquake therefore suggests that compressional tectonics related to subduction are not occurring in the Calabria regton. Its mechanism is different from those of the nearby Sicilian events, and so it cannot be interpreted as directly reflecting the African-Eurasian collision. As in Albania and NW Greece, the boundary between this area of normal faulting zyxw zyx zyxwvutsrq Active tectonics of the Adriatic region 969 and the nearby thrusting is very sharp and probably marks an important change in stress orientation. 5.8 SUMMARY 6 Motion of the Adriatic block The normal faulting in peninsular Italy cannot be explained in terms of the Africa-Eurasia convergence. As shown in Fig. 2 , the slip vector azimuths in central Italy should be NNW if they represent movement about the Africa-Eurasia pole. Similarly, if thrusting in Yugoslavia were related to the Africa-Eurasia motion, then the slip vector should also be NNW: approximately orthogonal t o that which is observed. Fig. 18 shows a summary of the slip vector directions derived from the well constrained focal mechanisms discussed in this paper, together with those predicted for the Africa-Eurasia convergence. Several groups of slip-vector orientations can be recognized. The N-S slip vectors observed in Sicily match reasonably those predicted for the Africa-Eurasia convergence. Those south of Kefallinia are related to subduction in the Hellenic Trench. The slip vectors of normal faulting earthquakes in Albania and NW Greece are not shown in Fig. 18, but are presumably related to the extension in the Aegean. Slip vector azimuths along the Yugoslavian and Albanian coasts appear to change in the Montenegro area. Perhaps the most striking feature of the seismicity in this region is the lack of activity in the Adriatic Sea itself. A few small. scattered earthquakes do occur in the Adriatic, particularly near the Gargano Ridge (Figs 3 and 4), but the activity is very much less intense than in the surrounding mountain belts: a feature that is reflected in the bathymetry of the Adriatic, w h c h is relatively flat. The larger earthquakes occur almost entirely in coastal or land areas, as seen from the locations of the earthquakes for which teleseismic first-motion fault-plane solutions are available (Fig. 5); which necessarily have magnitudes greater than about mb = 5.2. This pattern is also seen in the historical seismicity of the region (Anderson 1985). We believe that the absence, or low levels, of deformation in the Adriatic Sea indicate that its behaviour is that of a relatively rigid block within the deforming region, similar t o that of Central Turkey. the southern Caspian Sea and central Iran further east (Jackson & McKenzie 1984). It should therefore be possible to describe its motion relative to the Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 The fault-plane solutions in the circum-Adriatic regton show a clear pattern. West of Messina, thrust faulting occurs with slip vectors that appear t o reflect the overall motion between Africa and Eurasia (Fig. 5). In peninsular Italy only normal faulting mechanisms are found, but these change to pure dip-slip thrusting at the northern end of the Adriatic Sea. Inland northern Yugoslavia shows strike-slip and thrusting mechanisms, and inland strike-slip activity is also seen in southern Yugoslavia. Along the southern Dalmatian coast thrusting occurs on shallow landward-dipping faults. This thrusting continues south along the coast of Albania and northwestern Greece as far as Kefallinia. In this region, the bathymetry, the lack of intermediate depth earthquakes further north, and two right-lateral strike-slip mechanisms suggest that thrusting in the Hellenic subduction zone terminates against a strike-slip fault. South of Kefallinia, thrust faulting associated with subduction of the Mediterranean beneath the Aegean plate dominates the seismicity. The normal faulting in the Aegean extends slightly west of the highest topography in the Pindos mountains into Albania and NW Greece. The normal faulting in the 1908 Messina earthquake is not obviously related t o either the Africa-Eurasia convergence, or to active subduction in the Tyrrhenian Benioff Zone, or to the normal faulting observed elsewhere in peninsular Italy. zyxwvutsr 970 zyxwvutsrqp zyxwvu H. Anderson and J. Jackson Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxwvutsr zyxwvutsrqp Figure 18. Slip vectors (solid lines) derived from fault plane solutions shown in Fig. 5 . Only those slip vectors that are reasonably well defined b y the fault plane solution are shown. Dotted lines indicate the azimuth of slip vectors expected for movement about the African-Eurasian pole. In central Italy and Yugoslavia the observed slip vectors are at a high angle to the Africa-Eurasia direction and their orientations define an Adria-Eurasia pole of rotation at about 45.8"N, 10.2"E in northern Italy. Shaded areas indicate the seismically active borders of Adria which connect with t h e deformation zone between Africa and Eurasia in northern Sicily. The location of the Adria-Africa boundary is uncertain but the presence of some seismicity in the Strait of Otranto (Fig. 3) suggests that it may occur in this region. The more easterly azimuth of slip vectors along the Albanian coast and in the Messina Strait may indicate deformation associated with the Africa-Adria boundary but this cannot be conclusively demonstrated. Eurasian plate by a rotation about a n Eulerian pole, and we will now use the slip vectors of earthquakes in the deforming belts surrounding the Adriatic to help define such a pole. It is important to appreciate the reasons for undertaking such an exercise, which are (i) t o see whether rotation about a pole can account in a general way for the change in style and orientation of faulting around the boundaries of the Adriatic Sea, and (ii) t o use the pole to predict the overall rate and orientation of motion across the 100-200 km wide deforming belts that surround the Adriatic Sea. We d o not believe that the slip vectors on all the faults in these wide deforming zones will reflect the motion predicted by the pole of rotation, in zyxw zyx zyxw Active tectonics of the Adriatic region 971 Figure 19. Comparison of observed slip vectors from fault plane solutions (dotted lines) and slip vectors calcuhted from a pole of rotation at 45.8"N 10.2"E between Adria and Eurasia ( d i d lines). Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 the way expected within the narrower zones of deformation that bound plates in the oceans. Faulting in the wider deforming zones that bound relatively rigid continental blocks is distributed, may involve complicated geometries, and the overall resultant deformation is best described by a continuum approach (see McKenzie & Jackson 1983; Jackson & McKenzie 1984; Walcott 1984). However, in order to predict features of the continuum deformation, such as crustal thickening and palaeomagnetic rotations, and in order to investigate how faulting is able to take up the distributed deformation, the velocity boundary conditions across the deforming zone must be known (see e.g. Walcott 1984; McKenzie & Jackson 1986); these can usefully be predicted by the relative motion of the stable blocks on either side, which is described by a rotation about a pole. The most important constraint on the location of the Eurasia-Adria pole is the slip vector azimuth of the Friuli events. This direction is identical to that predicted by AfricaEurasia convergence (Fig. 18), and might imply a continuity of the Adriatic block with Africa (as a 'promontory'). However, the deforming margins of this block, in Italy and Yugoslavia, should then be deforming with the NNW slip vectors shown by the dashed lines in Fig. 18. They are not. Since the belt of seismicity surrounding the Adriatic is continuous, then the Friuli events cannot reflect African-Eurasia convergence and must be part of the Adria-Eurasia motion, An instantaneous pole of rotation has therefore been calculated using the Friuli slip vectors and the slip vectors of other central Italian and coastal Yugoslavian events. The observed and calculated slip vector azimuths for the best fitting pole of rotation, located in zy Ad Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxwvuts zyxwv zyxw zyx zy H. Anderson and J. Jackson 972 northern Italy at 45.8"N, 10.2"E, are shown in Fig. 19. Calculation of this pole did not include the four events in the Pannonian Basin (nos 8, 20, 21, 43) or the two events in Macedonia (6, 40) which may not be representative of the Adria-Eurasia motion. However, inclusion of these six events makes little difference; the resulting pole being at 46.O"N 10.2"E. Inclusion of the coastal Albanian events (17, 44) gives a pole of 46.2"N. 10.4"E, but the misfits t o these slip vectors are between 20" and 30", and it is unlikely that they occurred on part of the deforming Adriatic margin. The obvious question then is: where are the boundaries of Adria? The wide belt of seismicity running through Italy round the Alps and down through the Dinarides obviously marks the deforming edges of the Adriatic rigid block which is rotating anticlockwise about a pole in northern Italy. This seismicity marks the boundary with Eurasia but there is no evidence for the location of the Adria-Africa boundary. It must lie between Sicily and the Campania region of peninsular Italy, and probably crosses the Balkan coastline south of Montenegro. Although some historical seismicity is known from the Gargano Peninsular region (41.5"N 15.8"E), such as the moderate earthquake of 1627 (Molin & Margottini 1984), the southern part of the Apulia-Gargano region appears relatively aseismic and seems to be part of the stable Adriatic zone, so the Africa-Adria boundary probably lies t o t h e south of this region. One possible boundary zone is indicated i n Fig. 18. Very few earthquakes have been located in this zone but a small cluster of events in the Strait of Otranto may be important. None of these events was large enough for a fault-plane solution t o be determined, but they may mark the site of future important events. Since the Africa-Adria boundary is not marked by intense seismic activity it seems probable that the relative motion between the African plate and the Adriatic block is small. T h e relative motion between Africa, Eurasia and Adria can be compared at the Strait of Otranto. Fig. 20 shows the velocity triangle for this area in which Africa is moving NNW relative t o Eurasia at a rate of 8.8 mmyr-'. Adria is moving in a NE direction relative t o Eurasia at an unknown rate. If the relative motion between Africa and Adria is to be zyxwvutsrq zyxwvu Eur Figure 20. Velocity triangles for Africa (Afr), Eurasia (Eur), and Adria (Adr) at the Strait of Otranto. Velocities are shown in cm yr-'. The Adria-Africa motion is uncertain because only the direction and not t h e magnitude of the velocity of Adria with respect t o Eurasia is known. Possible relative motions between Adria and Africa are shown as dashed lines. The Africa-Eurasia motion is determined from rotation about the pole shown in Fig. 2 with an angular velocity of 1.42X10-7degyr-', derived by Chase ( 1 978). zyxwv zyx zyxwv Active tec'tcwiics o f the Adriatic region 973 minimized. then Adria moves SE relative to Africa a t a rate of about 7.3 nim yr-'. This would require an Adi-ia-Eurasia velocity of 5.0 mni yr-I. An alternative possibility is that. because the Adria-Eurasia deforming zone is the most seismically active, the velocity of Adria relative to Eurasia is greater than both the Africa -Eurasia and the Adria-Africa rates. Increasing tlie ad ria^- Eurasia velocity to 12.7 iiim yr-' predicts an E--W relative motion between Adria and Africa in this area. which is similar to the slip vector in the Messina earthquake: and thus the paucity of seismicity might reflect a major strain release associated with this earthquake. There is. however. no way to define the motion of Africa relative to Adria unless either the rotation rate o f the Adria -Eurasia motion 01- a focal mechanism for an earthquake clearly related t o the Africa--Adria motion can be determined. 7 Rates of deformation In this section we use Kostrov's (1074) result that the average tensor strain rate Ei; across a deforming zone can be obtained by summing the seistiiic-inornetit tensor elementsMi; of N earthquakes within the zone. Thus Mi;= M , ) ( U i ' 2 ; t- u; n i ) , where Mi;is the moment tensor, M , the scalar moment. and C and 6 ai-e unit vectors in the direction of the slip vector and the normal to the fault plane, respectively (see Aki & Richards 1980). The coordinate frame used was with the y = north, x = east and z = u p directions positive. For many of these earthquakes, the scalar moment could only be determined from a moment-magnitude relationship. Dziewonski & Woodhouse ( 1983) determined a relationship between surface wave magnitude and moment of: M, = 0.668 logMo - 10.86, where M, is in dyne -- cm (1 dyne - cm = 1 0-7 Nm). The errot-s in the calculation of surface-wave magnitude and i n the moment-magnitude relationship are difficult to assess. The equivalent moment values for surface wave magnitudes of 5.8 and 6.0 are 8.7 x 1017Nm and 1.7 x 10l8 Nm, respectively, so an error o f 0.2 in the M , value determined in this range produces a factoi- of two difference in the moment value. Because the relation is logarithniic, the errors in tlie moment assessment o f the thirteen largest earthquakes in the Adriatic seismic belt are probably much greater than the sum total of all the smaller events. Therefore. only the largest events were considered in our calculations. The moment tensor elements of these events are shown i n Table 4. The summed seismic-moment tensor elements for central Italy and coastal Yugoslavia are also shown in Table 4. The magnitudes of the individual moment tensor elements are very similar in these t w o areas, which they should be if central Italy and coastal Yugoslavia are opposing deforming margins of the relatively aseismic Adria. A similar result was obtained in a study of the twentieth century seismicity of the region by Anderson ( 1085). She used a moment-niagnitude relation to estimate static seismic moment (M,) rates for central Italy and Yugoslavia of I S and 27 x 1 0 1 7 N n 1yr-I. Table 4 also shows the summed moment Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxwvutsr zyxwvu zyxwvu zyxwvutsrqponmlkjih zyxwvutsrqponm zyx zyxwvuts where p is the rigidity, V the volume of the deforming zone. M;; is the moment tensor of the n t h earthquake, and r the time interval over which the moment tensors are summed. We obtained the six elements o f the seismic moment tensor for each first motion fault plane solution using the relation 974 zyxwvutsrqp zyxwvutsrq H. Anderson and J. Jackson zyxwvutsrqp Table 4, Seismic-moment tensor elements. Central Italy Event no. Date 23 47 38 48 49 5 42 7 1.07.15 84.04.29 79.09.19 84.05.07 84.05.11 62.08.21 80.1 1.23 Mo (X Mxx Myy MZZ lo'* Nm) 0.1 0.3 0.7 0.8 0.2 0.7 26.0 Sum: MXY Mxz Myz 0.0 0.1 0.7 0.5 0.1 0.4 7.4 0.0 0.1 -0.4 0.1 0.1 0.1 13.9 0 .o -0.2 -0.3 -0.6 -0.2 -0.5 -21.2 0.1 0.1 0.1 0.3 0.1 0.3 10.9 -0.1 -0.3 0.0 -0.2 -11.9 0.1 -0.2 -0.4 -0.5 0 .o -0.4 -8.2 9.2 13.8 -23.0 11.8 -12.6 -9.6 0.0 -0.1 Coastal Yugoslavia 62.01.1 1 84.05.1 3 79.04.15 79.05.24 79.04. I 5 79.04.09 6 .O 0.2 46.0 2.2 0.6 0.1 zyxwvut zyxwvu zyxwvut -3.8 0.0 -8.5 -1.1 0.0 0.0 -13.5 Sum 2.7 -0.1 -18.5 -0.3 0.0 0.0 1. I 0.1 27.0 1.4 0.0 -16.1 29.6 0 .o -3.9 0 .o -12.6 -0.6 0 .o 0 .o -2.5 0.0 -18.7 -1.5 -0.5 -0.1 1.7 -0.2 -32.1 -0.7 -0.3 0 .o -17.2 -23.4 -31.7 Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 4 50 35 37 36 34 Moment tensors rotated t o the frame x = azimuth 122". y = azimuth 032", z = vertical (positive): -23.0 Italy Yugoslavia tensors for Italy and Yugoslavia rotated into a new coordinate frame, in which the axes x l , x2, x 3 are positive in the directions 122", 032", and vertical. In this frame the x 2 direction is normal t o the strike of the deforming belt in central Italy. The rate of seismic extension across this zone is given by (see Kostrov 1974; Jackson & McKenzie 1987), where I is the length of the deforming zone along strike and t is its seismogenic thickness. For central Italy, taking I = 420 km, t = 15 k m , p = 3 x 10'0Nm-2 and = 21 yr gves an estimated extension velocity of 2.9 mm yr-'. This velocity normal t o the strike of the deforming zone (in azimuth 032") is the only one predicted by the motion of the 'rigid' blocks either side (see Jackson & McKenzie 1987). However, the pole of rotation between Adria and Eurasia found in section 6 predicts an overall direction of motion across the zone in central Italy of 058". Thus, t o have a resolved component in the direction 032" of 2.9mmyr-', the overall motion between Adria and Eurasia would have a magnitude of 3.2 mm yr-'. In coastal Yugoslavia, using values of 1 = 2 5 0 k m , t = 15 km and r = 21 yr, the seismic shortening normal to the zone (azimuth 032"), which is similar in direction t o the overall Adria-Eurasia motion predicted by the pole of rotation in Section 6 (azimuth 034") is calculated t o be 6.5 mm yr-'. Both these rates are uncertain by at least a factor of two, are dominated by the contribution of the largest earthquakes in the 21 yr period concerned (no. 42 in Italy and no. 35 in Yugoslavia), and ignore the unknown contribution of aseismic creep to the ', zy zyxw zyx zyxwvu zyxwv Active tectonics of the Adriatic region 975 8 Discussion It is now possible to demonstrate that the current deformation in the Adriatic area is not simply reflecting the N-S shortening between Africa and Eurasia. This was suggested by McKenzie (1972), but there were insufficient good focal mechanisms to demonstrate it conclusively at that time. The slip vectors of 24 fault-plane solutions show that the bulk movement of the relatively stable Adriatic block can be approximated by an anticlockwise rotation relative to Eurasia about an Eulerian pole at 45.8ON, 10.2"E. The current extensional deformation in Italy and the thrusting in northern Italy and Yugoslavia therefore reflect the deforming margins of the relatively aseismic Adriatic area. An alternative view might be that the Adriatic Sea is still attached t o Africa in some sense but deforms internally: slip vectors within the deforming region would then not be required t o reflect the overall Africa-Eurasia motion. We feel this view is unnecessarily complicated and does not explain why (a) the observed slip vectors match those predicted by the rotation about a pole in N. Italy (Fig. 19), and (b) the seismicity is concentrated in the land and coastal areas, which are generally mountainous, while the sea-floor is relatively flat and aseismic. This alternative view requires that the Adriatic Sea deforms dramatically by creep and yet produces n o substantial topography or bathymetry. It also begs the question: to what extent is it meaningful t o consider the Adriatic Sea as a promontory of Africa if it is not rigid but deforms internally? We feel that the description we offer, in which the Adriatic Sea acts as an effectively rigid block rotating relative t o Eurasia about a pole in N. Italy, describes the general nature of the deformation in the wide deforming belts separating the Adriatic and Eurasia in a far simpler way. Unlike many of the other western Mediterranean basins, the Adriatic Sea area is underlain by continental crust that is typically 25-36 km thick and reaches a maximum of 35-40 km thickness beneath the southern Adriatic basin (Dragasevic 1973; Nicholich 1981). An important lithospheric discontinuity in the Gargano Ridge area is also suggested by Calcagnile & Panza (1981). From inversion of the regional dispersion relations derived from surface waves, combined with the results of crustal refraction surveys, they recognize a change in the lithospheric thickness from about 70 k m in the northern Adriatic basin to more than 110 km in the southern Adriatic basin. A scattered band of weak seismic activity crosses the central Adriatic Sea (Fig. 3) near the Gargano Ridge. The change in crustal and lithospheric structure in this area may mark some internal deformation of the Adriatic Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 overall deformation, However, the rates calculated using a 21 yr seismicity are similar in magnitude to those found by Jackson & McKenzie (1987) using a 70 yr seismicity: they estimated seismic extension rates of 1.3-3.5 mm yr-' in central itaiy and seismic shortening of 1 .O-2.4 mm yr-' in coastal Yugoslavia. Predicted velocities between Adria and Eurasia in the strait of Otranto are roughly double those in central Italy (because of the increased distance from the pole of rotation). The Adria-Eurasia velocities estimated from the seismicity above are similar in magnitude to those shown in the tentative velocity triangles for the Strait of Otranto in Fig. 20. However, their considerable uncertainty, combined with the unknown contribution of aseismic creep, does not allow them t o help in solving the enigma of the Africa-Adria boundary. This boundary is not defined by a belt of intense seismicity, so its nature and location are uncertain (Fig. 18). The slip vector in the Messina earthquake (Fig. 5) was approximately E-W, and may represent deformation in the Adria-Africa boundary zone. However, such a suggestion is very tentative and must await further evidence in the form of large earthquakes in the Ionian Sea. Calabria or Strait of Otranto. zyxwvutsrq zyxw zyxwvutsrq zyx zyxwvutsr 976 H. Andersori and J. Jacksori 9 Conclusions This study demonstrates that in the Adriatic region, as elsewhere in the Alpine-Himalayan seismic belt (e.g. Molnar & Tapponnier 1981; Jackson & McKenzie lc)X4), there is a large continental block in which the seismicity is relatively feeble. The relative rigidity of this block allows its bulk motion to he described by rotation about a pole, and goes some way towards accounting for the differences in slip vector directions, deformation style and levels of seismicity around its boundaries. The N--S shortening in Sicily changes t o N E --SW extension in peninsular Italy, which in turn changes to thrusting in N. Italy. A belt of active shortening exists from N. Italy southwards along the Yugoslavian and Albanian coasts, and into Greece. The superior data of the last 21 years support the suggestion from all the 20th Century seismicity; that the extension rate in southern peninsular Italy and the shortening rate in southern coastal Yugoslavia are about equal. The seismicity accounts for about 2.0 nimyr-' of motion in these regions. but the velocity may be much greater if aseismic creep contributes substantially t o the deformation. Thus the Adriatic Sea is surrounded by belts o f high topography and seismicity that are about 100-150 km wide. The new set of fault-plane solutions presented here strongly support the suggestion that the Adriatic region has become detached from the African continent. The boundary between Adria and Africa is not marked by a belt of intense seismicity but may be located in the southern Adriatic Sea, near the Strait of Otranto. Acknowledgments This work was supported in part b y a grant from the Natural Environment Research Council. H.J.A. acknowledges a postgraduate funding award from the New Zealand National Research Advisory Council. We thank Dan McKenzie and Rob Westaway for many helpful discussions, and a reviewer for useful comments on the manuscript. H. Campbell helped with draughting. This is Cambridge University Earth Sciences Contribution No. 987. References Academy of Sciences, Albania, 1983. The earthquake of 1979 April 15 and the elimination of its consequences, Reports and papers on the Shkodra symposium, 1980, Nentori Publishing House, Tirana. Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 block; possibly a relic of the former behaviour of the Adriatic block as a 'promontory of Africa'. Recent interpretations of palaeoinagnetic data by Vandenberg ( 1983) and Vandenberg & Zijderveld ( 1981) suggest that the Italian peninsula, Sardinia and the southern Alps formed a single 'Adriatic continental block' which rotated twice: once during the Late Cretaceous as a northern promontory of the African plate, and for a second time, independently from Africa and Europe, in the post-Early Tertiary. The focal mechanisms presented here show that this recent anticlockwise rotation, which has amounted t o about 30" since the Early Tertiary, is still continuing. Finally. we should stress that the kinematics discussed in this paper, because they are based o n earthquake mechanisms, are restricted solely t o a description of present-day motions. These need not be representative of longer term motions and the kinematics described here are unlikely t o he the same as those prior to the Pliocene. The pre-Pliocene deformation of this area is certainly complicated (see Mantovani, Babbucci & Farsi 1985, among others) and probably involved motions different from those occurring today. zyx zyxwvutsrqp zyxwvuts Active tectonics of the Adriatic region 977 Aki, K. & Richards, P. G., 1980. Quantitative Seismology, 2 volumes, W. H. Freeman & Co., San Francisco. Aliaj, S., 1982. Basic seismotectonic features of Albania, in Earthquake Risk Reduction in the Balkan Region. Working Group A , Seismology, Seismotectonics, Seismic Hazard and Earthquake Prediction, Final Report, A27-A31. Ambraseys, N. N.. 1966. Seiamic encironrnent of the Skopje earthquake of J u l y , 1963, Rev. d e I'Union int. d e Secours, 5 , 1 -1 9. Ambraseys, N. N., 1975. Studies in historical seismicity and tectonics, in Geodynamics T o d a y , pp. 7-16, Royal Society, London. Ambraseys, N. N. & Melville, C. P., 1982. A History of Persian Earthquakes, Cambridge University Press. Ambraseys, N. N. & Morgenstern, N., 1966. Field report on the Skopje earthquake of July 26, 1963, Department of Civil Engineering, Imperial College of Science & Technology, London. Anderson, H. J., 1985. Seismotectonics of the western Mediterranean, PhD thesis, University of Ca mb r idg e . Argand, E., 1924. La tectonique d e I'Asie, Proc. Int. Geol. Congr., XIII, 17 1-372. Arsovski, M., 1970. Contemporary tectonic properties of some seismic active zones in Yugoslavia, Bulgarian Acad. Sci., Earthquake Engineering, Proc. 3rd Europ. Symp. Earthquake Engineering, Sofia, 1970, 181-188. Arsovski, M. & Iladzievski. D.. 1970. Correlation between neotcctonics and acismicity o f Macedonia. Tectonophys., 9 , 129-142. Arsovski, M., Gojgic, D., Griljic, N., Mitrev, T. & Hadzievski, D., 1966. Summary of the seismo-geologic investigation in the Skopje valley, Univ. Skopje, Inst. of Seismology, Earthquake Engineering and Town Planning, 1. Barazangi, M. & Dorman, J., 1969. World seismicity map from ESSA, Coast and Geodetic Survey, 1961-1967. Bull. seism. SOC.A m . , 59,369-380. Barbano, M. S., Cosentino, M., Lombardo, G. & Patane, G., 1980. Isoseismal maps of Calabria and Sicily earthquakes (southern Italy). Consiglio nazionale delle richerche progetto finalizzato geodinamica, Puhl. no. 3 4 1 , Univ. di Catania, Italy. Berberian, M., Jackson, J. A., Ghorashi, M. & Kadjar, M. H., 1984. Field and teleseismic observations of the 1981 Golbaf-Sirch earthquakes in SE Iran, Geophys. J . R . astv. Soc., 77,809-838. Berg, G. V., 1964. The Skopje, Yugoslavia, earthquake, July 2 6 , 1 9 6 3 , American Iron and Steel Institute. Boore, D. M., Sims, J. D., Kanamori, H. & Harding, S., 1981. The Montenegro, Yugoslavia earthquake of April 15, 1979: source orientation and strength,Phys. Earth planet. Int., 27, 133-142. Brooks, M. & Ferentinos, G., 1984. Tectonics and sedimentation in the Gulf of Corinth and the Zakynthos and Kefalltnia Channels, western Greece, Tectonophys., 101,25-54. Cagnetti, V. & Pasquale, V., 1979. The earthquake sequence in Friuli, Italy, 1976, Bull. seism. SOC.A m . , 69(6), 1797-1818. Calcagnile, G. & Panza, G . F., 1982. The main characteristics of t h e lithosphere-asthenosphere system in Italy and surrounding regions, Pure nppl. Geophys., 119,865-879. Celet, P., 1977. The Dinaric and Aegean arcs: the geology of the Adriatic, in The Ocean Basins and Margins, 4 A , pp. 215-261, eds Nairn, A. E. M., Kanes, W. H. & Stehli, F. G. Channell, J. E. T., D'Argenio, B. & Horvath, F., 1979. Adria, the African promontory, in Mesozoic Mediterranean palaeogeography, Earth Sci. Rev., 15,213-292. Chase, C. G., 1978. Plate kinematics: the Americans, East Africa and the rest of the world, Earth planet Sci. Lett., 31, 355-368. Christensen, D. H. & Ruff, L. J., 1985. Analysis of the trade-off between hypocentral depth and source time function,Bull. seism. Soc. A m . , 7 5 , 1637 1656. Cipar, J., 1980. Teleseismic observations of the 1976 Friuli, Italy earthquake sequence, Bull. seism. Soc. A m . , 70, 963-983. Cipar, J., 1981. Broadband time domain modelling of earthquakes from Friuli, Italy, Bull. seism. SOC. A m . , 71, 1215-1231. Console, R. & Favali, P., 1981. Study of the Montenegro earthquake sequence (March-July 1979), Bull. seism. SOC.A m . , 71,1233-1248. Cosentino, P. & Mulone, A., 1985. The Belice earthquake of January 15, 1968, in Atlas of isoseismal maps of Italian earthquakes, ed. Postpischl, D., Consiglio Nazionale delle Richerche, Roma. Cvijanovic, D. & Prelogovic, E., 1977. Seismicity and neotectonic movements of the Croatian region (SFR Yugoslavia), Publ. Inst. Geophys. Pol. Acad. Sci., A5(116), 281-290. Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyx zyxwv zyxwvutsrqp zyxwv zyxwvutsr 978 zyxwvutsrqp zyxwvutsr zyxwvut zyxwvuts zyxwvutsrq zy H. Anderson and J . Jackson Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 D’Argenio, B. & Horvath, F., 1984. Some remarks on the deformation history of Adria, from the Mesozoic to the Tertiary, Ann. Geophys., 2(2), 143-146. D’Argenio, B., Horvath, F. & Channell, J. E. T., 1980. Palaeotectonic evolution of Adria, the African promontory, in Gdologie des Choinesalpines issues de la Tethys, eds Aubouin, J . , Debelmas, J. & Latreille, M., M i m . Bur. Rech. Geol. Min., 115, 331 -351. de Panfilis, M. & Marcelli, L., 1968. I1 period0 sismico della Sicilia occidentale inizato il 14 Gennaio 1968, Annls. Geofis., XXI,343-421. del Pezzo, E. & Martini, M., 1982. The April 15, 1978 event of the Gulf of Patti, Sicily, and its aftershocks sequence, Boll. Geof teor. oppl., XXIV(94), 109-119. del Pezzo, E., Maresca, R., Martini, M. & Scarpa, R., 1984. Seismicity of the Aeolian islands, southern Italy, Ann. Geophys., 2(2), 173-180. Deschamps, A. & King, G. C. P., 1983. The Campania-Lucania (southern Italy) earthquake of 23 November, 1980, Earth planet. Sci. Lett., 62, 296-304. Deschamps, A. & King, G. C. P., 1984. Aftershocks of the Campania-Lucania (Italy) earthquake of 23 November 1980,Bull. seism. Soc. Am., 74,2483-2517. Deschamps, A,, Iannaccone, G. & Scarpa, R., 1984. The Umbrian earthquake (Italy) of 19 September, 1979, Ann. Geophys., 2(1), 29-36. Di Filippo, D. & Peronaci, F., 1962. Contributo alla tectonica delle Alpi Dinariche, Annls. Geofis., 15, 379-391. Dragasevic, T., 1973. Savremena grada Zemljine kore i gornjeg omotaca na podrucju Jugoslavije, Vesnik, Knj., XIV/XV,Ser. C, 41-52. Dziewonski, A. M. & Woodhouse, J. H., 1983. Studies of the seismic source using normal-mode theory, in Earthquakes: Observations Theory and Interpretation, pp. 45-1 37, eds, Kanamori, H. & Boschi, E., Italian Physical Society, North Holland Publishing Co., Amsterdam. Dziewonski, A. M., Chou, T. A. & Woodhouse, S. H., 1981. Determination of earthquake source parameters from waveform data for studies of global and regional seismicity, J. geophys. Res., 86, 2825-2852. Dziewonski, A. M., Friedman, A. & Woodhouse, J. H., 1983. Centroid-moment tensor solutions for January-March, 1983, Phys. Earth planet. Int., 33,71-75. Dziewonski, A. M., Franzen, J. E. & Woodhouse, J. H., 1985. Centroid-moment tensor solutions for April-June, 1984, Phys. Earth planet, Int., 37,87-96. Fitch, T. J. & Mukhead, K. J., 1974. Depths to larger earthquakes associated with crustal loading, Geophys. J. R . ash. Soc., 37, 285-296. Gasparini, C., Gasperini, M., Iannaccone, G., Napoleone, G., Scarpa, R., Stucchi, M., Taccetti, Q. & Zonno, G., 1980. Seismometrical observations: elaboration and preliminary interpretation of the Norcia earthquake, 1979, Annls. Geofis., 33, 101-119. Gasparini, C., Iannaccone, G., Scandone, P. & Scarpa, R., 1982. Seismotectonics of the Calabrian arc, Tectonophys., 84,267-286. Gasparini, C . , Iannaccone, G. & Scarpa, R., 1985. Fault-plane solutions and seismicity of the Italian peninsula, Tectonophys., 117, 59-78. Ghisetti, F., Scarpa, R. & Vezzani, L., 1982. Seismic activity, deep structures and deformation processes in the Calabrian arc, southern Italy, Earth Evol. Sci., 3, 248-260. Giardini, D., Dziewonski, A. M., Woodhouse, J . H. & Boschi, E., 1984. Systematic analysis of the seismicity of the Mediterranean region using the centroid-moment tensor method, Bull. Geofis. teor. appl., XXVI(103), 121-142. Giese, P. & Reutter, K.-J., 1978. Crustal and structural features of the margins of the Adria microplate, in Alps, Apennines, Hellenides, pp. 565-588, eds Closs, H., Roeder, D. & Schmidt, K., Schweizerbart, Stuttgart. Hsu, K. J., 1982. Alpine Mediterranean geodynamics: past, present and future, in Alpine Mediterranean Geodynamics, eds Berckhemer, H. & Hsu, K. J., A m . Geophys. Un. Geodyn. ser., 7, 7-14. Hurtig, E. & Neunhofer, H., 1980. Seismologische expedition des zentrolinstituts fur Physik der Erde der AdW in das Herdgebiet des Erdbebens vom 1.5 April 1979 in Montenegro (SFR Jugoslawien), Zentralinstitut Fur Physik der Erde, Potsdam. I.F.P., 1966. Etude GPologique de I’Epire (Grece nord-occidentale), 2 vols., Paris Editions Technip (L’lnstitut de Geologie et Recherches du sous-sol, Athenes, et Institut Francais du Petrole). Irby, W. L., Jacobs, W. S., Minsch, J. H., Needham, R. E., Person, W. J., Presgrave, B. W. & Schmieder, W. H., 1985a. Preliminary determinotion of epicenters, September 1984, Monthly Listing, National Earthquake Information Service, U.S. Department of the Interior, Washington. zyxwvuts zyxwv zyxwv zyx zyxwvutsrq zyxw Active tectonics of the Adriatic region 979 Irby, W. L., Jacobs, W. S., Minsch, J. H., Needham, R. H., Person, W. J., Presgrave, B. W. & Schmieder, W. H., 1985b. Preliminary determination of epicenters, July 1984, Monthly Listing, National Earthquake Information Service, U.S. Department of the Interior, Washington. Jackson, J. A., 1979. Active faulting and continental deformation, PhD thesis, University of Cambridge. Jackson, J. A. & Fitch, T. J., 1979. Seismotectonic implications of relocated aftershock sequences in Iran and Turkey, Geophys. J. R . astr. Soc., 57,209-229. Jackson, J. A. & McKenzie, D. P., 1984. Active tectonics of the Alpine-Himalayan belt between western Turkey and Pakistan, Geophys. J. R . astr. SOC.,77,185-264. Jackson, J. A. & McKenzie, D. P., 1988. The relationship between plate motions and seismic moment tensors, and the rates of active deformation in the Mediterranean and Middle East, Geophys. J. R. astr. Soc.. in press. Jackson, J . A , , Gagnepain, J.. Houseman. G., King, G. C. P., Papadimitriou, P., Soufleris, C. & Virieux, J . . 1982a. Seismicity, normal faulting, and the geomorphological development of the Gulf of Corinth (Greece): the Corinth earthquakes of February and March 1981, Earth planet. Sci. Lett., 57, 377-397. Jackson, J. A,, King, G. & Vita-Finzi, C., 1982b. The neotectonics of the Aegean: an alternative view, Earrh plunet. Sci. Lett., 61, 303-318. KadinskyCade, K. & Barazangi, M., 1982. Seismotectonics of southern Iran: the Oman line, Tectonics, 1(5), 389-412. Kanamori, H. & Given, J. W., 1981. Use of long-period surface waves for rapid determination of earthquake-source parameters, Phys. Earth planet. Int., 27, 8-31. King, G., Tselentis, A., Gomberg, J., Molnar, P., Roecker, S . , Sinvhal, H., Soufleris, C. & Stock, J., 1983. Microearthquake seismicity and active tectonics of northwestern Greece, Earth planet. Sci. Lett., 66,279-288. King, G., Ouyang, Z., Papadimitriou, P., Deschamps, A., Gagnepain, J., Houseman, G., Jackson, J., Soufleris, C. & Virieux, J., 1985. The evolution of the Gulf of Corinth (Greece): an aftershock study of the 1981 earthquakes, Geophys. J. R. astr. Soc., 80,677-693. Kociaj, S. & Sulst*arova, E., 1980. The earthquake of June 1, 1905. Shkodra, Albania; intensity distribution and macroseismic epicentre, Tectonophys., 67, 319-332, Kostrov, B. V., 1974. Seismic moment and energy of earthquakes, and seismic flow of rock, Izv. Acad. Sci. USSR Phys. Solid Earth, 1,23-40. Langston, C. A. & Franco-Spera, M., 1985. M o d e b g of the Koyna, India, aftershock of 12 December 1967, Bull. seims. SOC.A m . , 75(3), 651-661. Langston, C. A. & Helmberger, D. V., 1975. A procedure for modelling shallow dislocation sources, Geophys. J. R. astr. Soc., 42, 117-130. Lavecchia, G., Minelli, G. & Pialli, G., 1984. L'appennino umbro-marchigiano: tettonica distensiva e ipotesi di sismogenesi, Boll. Soc. geol. Ital., 103,467-476. Mantovani, E., Babbucci, D. & Farsi, F., 1985. Tertiary evolution of the Mediterranean region: major outstanding problems, Boll. Geofis. teor. appi., 27,67-90. McKenzie, D. P., 1972. Active tectonics of the Mediterranean Region, Geophys. J. R . astr. Soc., 30, 109-1 85. McKenzie, D. P., 1978. Active tectonics of the Alpine-Himalayan belt: the Aegean Sea and surrounding regions, Geophys. J . R . astr. Soc., 55, 217-254. McKenzie, D. P. & Jackson, J. A., 1983. The relationship between strain rates, crustal thickening, paleomagnetism, finite strain, and fault movements within a deforming zone, Earth planet. Sci. L e f t . , 65,182-202. McKenzie, D. P. & Jackson, J . A., 1986. A block model of distributed deformation by faulting, J. Geol. Soc. London, 143,349-353. Mercier, J . L., Carey, E., Philip, H. & Sorel, D., 1976. La neotectonique plio-quaternaire de l'arc eg6en externe et d e la mer Egee et ses relations avec la seismicite, Bull. soc. gbol. Fr., 7(XVIII), 355-372. Mercier, J. L., Delibassis, N., Gauthier, A., Jarrige, J.-J., Lemeille, F., Philip, H., Sebrier, M. & Sorel, D., 1979. La neotectonique de l'arc egeen, Rev. geol. syn. geogr. phys., 21.67-92. Minster, J. B. & Jordan, T. H., 1Y78. Resent day plate motions, J. geophys. Res., 83,5331 -5334. M o b , D. & Margottini, C., 1984. I1 terremoto del 1627 nella Capitanata Settentrionale, in Contributo Sismicita del territorio Italiano, Udine, 12 May 1981, pp. 251-279, Publs. Commissione ENEA-ENEL. Molnar, P. & Tapponnier, P., 1981. A possible dependence of tectonic strength on the age of the crust in Asia, Earth planet. Sci. Lett., 52, 107-1 14. Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 zyxwvut zy zyxwv 980 zyxwvuts zyxwv zy zyx zyxwvutsrqp zyxwvut zyxw zyxwvuts H. Anderson and J. Jackson Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 More&, C., 1984. Promontorio African0 o microplacca Adriatica, Boll. Ocean. teor. appl., 11(2), 151-168. Morclli. C.. Giew. P.. ('assinis. R.. C'olombi. 13.. Gucrra. I.. Luonpo. G.. Scarascia. S. & Schutte. K.-G., 1975. Crustal structure of southern Italy. A seismic refraction profile between Puglia CalabriaSicily, Boll. Geofis. teor. appl., XV11(67), 183-210. Mukrgia, F. & Boschi, E., 1982. The 1908 Messina earthquake and related seismicity, in Earthquakes: observation, theory and interpretation, pp. 493-518, eds Kanamori, H. & Boschi, E., Italian Physical Society, North Holland Publishing Co, Amsterdam. Nicholich, R., 1981. Crustal structures in the Italian peninsula and surrounding seas: a review of DSS data, in Sedimentary Basins of Mediterranearr Margins, pp. 3-17, ed. Wezel, F. C., C.N.R. Italian Project of Oceanography, Italy. Ouyed, M . , Yielding, G., Hatzfeld. D. & King, G . C. P., 1983. An aftershock study of the El Asnam (Algeria) earthquake of 1980 October 10, tieophys. J. R. astr. Soc., 73,605-639. Papazachos, B. C., 1973. Distribution of seismic foci in the Mediterranean and surrounding area and its tectonic implication, Geophys. J. R. astr. Soc., 33,421-430. Philip, H., 1983. La tectonique actuelle et recente dans le domaine mediterraneen et ses bordures, ses relations avec la sismicite, 71iese, Univ. des SL.\;,-[;.. ch. du Languedoc. Montpellier. Ritsema, A. R., 1975. The contribution of the study of seismicity and earthquake mechanisms to the knowledge of Mediterranean geodynamic processes. in Progress in tieodynamics, pp. 142- 153. Amsterdam. Ryan, W. B. F. & Heezen, B. C., 1965. Ionian sea submarine canyons and the 1908 Messina turbidity current, Bull. geol. Soc. A m . , 76, 915-932. Schick, R., 1977. Eine seismotektonische Bearbeitung des Erdbebens von Messina iin Jahre 1908, Geol. Jh., E-11, 3-74. Scott, D. R. & Kanamori, H., 1985. O n the consistency of moment tensor source mechanisms with firstmotion data,Phys. Earth planet, Int., 37, 97-107. Sorsky, A. A,, 1968. Geological and tectonic factors in the Skopje earthquake of 26 July 1963, in The Skopje earthquake of 26 July 1 Y63, Report of the Unesco Technical Assistance Mission, Unesco. Soufleris, C., 1980. The Thessaloniki (north Greece) 1978 earthquake sequence, PlzD rhesis, University of Cambridge. Soufleris, C. & Stewart, G., 1981. A hource study of the Thessaloniki (northern Greece) 1978 earthquake sequence, Geophys. J. R . astr. Soc., 67, 343-358. Soufleris, C., Jackson, J., King, G., Spencer, C. & Scholz, C., 1982. The 1978 earthquake sequence near Thessaloniki (northern Greece), Geophys. J. R . astr. Soc., 68,429-458. Stein, S., Wiens, D. A. & Fujita, K., 1982. The 1966 Kremasta reservoir earthquake sequence, Earth planet. Sci. Lett., 59,49-60. Suess, E., 1883.Das Antlitz der Erde, Tempsky (Prague) und Freytag (Leipzig). Sulstarova, E., 1980. Some characteristics of earthquake foci in Albania and the field of tectonic stress, Publ. Inst. tieophys. Pol. Acad. Sci., A-9(135), 79-86. Sulstarova, E., 1982. The focal mechanisms of earthquakes in Albania, in Earthquake Risk Reduction in the Balkan region, Working Group A , Seismology, Seismotectonics, Seismic Hazards and Earthquake Predictions, Final Report, A32-A38. Sulstarova, E. & Kociaj, S., 1980. The Dibra (Albania) earthquake of November 30, 1967, Tecronophys., 67,333-343. Sulstarova, E., Kociaj, S. & Aliaj, S., 1982. Seismic focal zones in Albania, indicated by seismological and geological data, in Earthquake Risk Reduction in the Balkan region. Working Group A , Seismology, Seismotectonics, Seismic Hazards and Earthquake Predictions, Final Report, A22-A25. Vandenberg, J., 1983. Reappraisal of paleomagnetic data from Gargano (South Italy), Tectonophys., 98, 29-41. Vandenberg, J. & Zijderveld, J. D. A., 1982. Paleornagnetism in the Mediterranean area, in Alpine Mediterranean tieodynamics, eds Berckhemer, H. & Hsu, K. J., Am, Geophys. Un. Geodyn. ser., 7, 83-1 12. Walcott, R. I . , 1984. The kinematics of the plate boundary zone through New Zealand: a comparison of short- and long-term deformations, tieophys. J. R. astr. Soc., 79, 613-633. Weber, C. & Courtot, P., 1978. Le seisme d u Frioul (Italie, 6 mai 1976) dans son contexte sismotectonique, Rev. Gdogr. phys. Gdol. dyn., 2(XX), 247-258. Westaway, R., 1987. The Campania, Southern Italy, earthquakes of 21 August 1962, Geophys. J. R. astr. SOC.,88, 1-24. zyx zyxwvutsrq zyxwvut zyxwvu zyxwvutsr Active tectonics ojthe Adriatic region 98 1 Westaway, R. W. C. & Jackson, J. A,, 1984. Surface faulting in the southern Italian CampaniaBasilicata earthquake of 23 November 1980, Nature, 312(5993), 436-438. Westaway, R. & Jackson, J., 1987. The earthquake of 2 3 November 1980 in Campania-Rasilicata (S. Italy), G'eophys. J. R. astr. Soc.. 90, 375-443. Yielding, G., Jackson, J. A , , King, G . C. P.. Sinvhal, H., Vita-Finzi, C. & Wood, R. M., 1981. Relation\ between surpdce deformation, fault geometry, seismicity and rupture characteristics during the El Asnam (Algeria) earthquake of 10 October 1980, Earth plaizet. Sci. L e f t . , 5 6 , 287-304. Zatopek, A,, 1968. The Skopje earthquake of 26 July 1963 and the seismicity of Macedonia, in The Skopje earthquake of' -36July IY63, Report of t h e Unesco Technical Assistance Mission, Unesco. Zonno, G. & Kind, R., 1984. Depth determination of North Italian earthquakes using Grafenberg data, Ritll. seistn. SOC.A m . , 74(5), 1 6 4 5 . ~1659. zyxwvutsr zyxwvut Appendix 1 KO . I. I"U I oh(; DLI'III '1, SrNS IIR i8. I0 lS..?S 10. 7. 1) 0 4 LO 00.0 38. I 1 20.72 (1 . 7.2 0 9 23 51.2 0.0 0 17 8 40.5 - 5.7 0 5 5 4.1 - zyxwvu 0. 3 37.80 20.56 4 43.30 1?.S 10. !I 41.02 lJ.98 8. 6 3L.100 LI.10C 5. i 7 . IUO 10.!lIlll 1;. 45.300 18.100 7 . 39.02 21.82 _>. -. 8. - - 2; I8 1?1 33.3 - .5.3 222 4 l i 12.5 - 5.6 45 !;1 1 .3 47 5b.3 8 30 3.6 109 OL so 02 01 i9 45.i 3.4 88 7 9 11.5 5.4 38.8 1 ?1.10 LO. 5.6 5.; 39. ,I i l l 2 1 . I! I 0 II. 5.6 41 .31 20.44 ?I. b.0 92 .?J7.750 12.983 10. 5.4 05 37.857 12.976 36. 5.1 71 3-.68; 12.966 3. 5.1 69 iT.900 z0.goo 6. $14 40. so0 19. 9(10 18. i.4 5.1 3 7 . 561 LO. 277 33. 39. 570 2O.63ll 8. 4J.8:.1 17.286 _. 23. .13.923 1'.?32 53. '1 I . (19s I7 . 7 I .1.I. 776 lO.335 38.233 LO. 540 ~ ~ - - zyxwvuts zyxwvuts zy 7 -.> 16 9 51.5 - I 8.5 - 42 14.3 - 56 48.7 - 3') 57.1 - 2.5.8 5.5 811 22 I2 5.4 97 8 9 17.5 5.4 5.6 106 1 2 28.5 5.0 5.5 81 15 36 51.8 5.6 2r , s- 11- 8 10 58.i 6.1 3.5. .;.2 Ill1 I 53.1 5.7 8. ;.2 !IS I 35 2Z.i - 5.b 114 I4 7 15.6 6.3 .. .I >. i8.89b ?O.'l-ll 8. 5.8 99 15 52 11.7 5.5 .4L1.556 13.X.5 9. 0.0 275 20 0 11.6 6.5 Y-. 56il 2 0 . 352 3.5. 5.8 202 16 59 ~ 1 8 . 2 b.4 -* 20.551 8. i.s 1611 0 59 tb.9 13.15; 16. 5.2 I20 16 il I2.O.;.s I ,3.5 46. ? X U Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 Loclition inft~rmation f o r focal ii~eclianisiiisfigures in this s t u d y . Latirude. longitude and deptli are derived mainly from USGS listings but soiiie informatioil is based on relative relocations. niacroseisinic and waveform modelling studies. All other data, such as inagnitudes. origin time and nuinber of stations reporting the event were deterinined f r o m USGS data 5.j 56.199 1 3.2 0.5 21). .5.3 114 16 75 3.3 5.1 46.302 13.19i 10. 5.7 25;. 3 15 19.9 6.0 46.522 15.132 .'I 5.4 2411 9 21 19.1 5.9 i 8 . 39 I Ii. OGb 21. 5.5 L11 23 .>.> _- .I^.? .5.- & zyxw zyxwvuts zyxwvutsr zyxwvutsrqpo H. Anderson and J. Jackson Appendix 1-continued "b DATE LAT LONG DEFlH HR MIN 34 790409 41.956 19.023 10. 5.3 196 2 35 790415 42.096 19.209 10. 6.2 217 6 36 790415 42.319 18.682 10. 5.7 291 37 790524 42.255 18.752 8. 5.8 No. STNS SFC "s 10 20.3 4.9 19 44.1 6.5 14 43 6.0 5.6 342 17 23 18.2 6.2 38 790919 42.812 13.061 16. 5.9 175 21 35 37.2 5.8 39 791208 38.284 11.741 33. 5.4 190 4 6 34.3 5.3 40 800518 43.294 20.837 9. 5.7 164 20 2 57.5 5.8 41 800528 i5.482 14.252 12. 5.7 232 19 51 19.3 5.5 42 801123 40.760 15.330 10. 6.0 265 18 34 53.8 6.9 43 810813 44.849 17.312 16. 5.4 143 2 58 11.9 5.5 44 821116 40.883 19.590 21. 5.6 241 23 41 21.0 5.5 830117 38.026 20.228 14. 6.1 329 12 41 29.7 7.0 830323 38.294 20.262 19. 5.8 258 23 51 6.5 6.2 47 840429 43.260 12.558 12. 5.2 202 0.0 5.3 48 840507 49 840511 50 840513 51 840709 5 3 Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016 45 46 zyxwvutsrq zyxwvu 41.765 13.898 10. 5.5 302 17 49 41.6 5.8 41.831 13.961 14. 5.2 259 10 41 49.9 5.2 42.967 17.734 30. 5.1 190 12 45 55.8 5.1 40.677 21.831 10. 5.1 208 18 57 09.6 4.9 zyxwv Appendix 2 Strike, dip and rake of events described in Appendix 1. Convention follows Aki & Richards (1 980). Rake was determined graphically. NODAL PLAW 1 STRIICI: 1 2 3 8 9 10 11 I? 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 208 163 310 197 310 303 064 302 252 204 197 200 270 250 270 120 164 145 152 250 172 115 148 180 DIP 55 34 83 80 65 74 64 55 66 70 56 58 50 58 64 71 40 88 83 88 90 71 84 80 44 75 80 70 73 80 40 68 55 80 NODAL PLANE 2 IWT STRIKE DIP 149 -110 -21 90 100 -100 113 -44 -80 35 18 31 65 97 90 61 12 0 90 132 -26 90 87 90 90 90 90 126 61 153 101 349 330 066 293 171 039 244 104 096 334 315 0 156 150 165 357 336 325 352 138 336 332 34 6 042 156 267 352 295 256 271 047 297 254 312 42 56 16 60 32 70 26 36 26 30 56 34 64 75 62 31 50 02 30 78 90 19 42 65 46 15 10 20 17 10 59 36 68 15 W 12 - 54 -163 90 76 - 69 43 -1.17 -106 134 147 150 141 84 90 166 178 180 90 9 -168 90 101 90 90 90 90 64 140 38 43 zyxw zyx zyxwvutsr zyxwvu zyxwvutsr zyxwvutsrqp Active tectonics of the Adriatic region 983 Appendix 2-continued N o w PLANE 1 EVENT No. 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 so 51 STRIKE DIP RAKE 121 148 154 320 254 294 052 317 165 297 135 027 72 90 70 66 56 83 62 62 60 3s 83 59 21 31 43 16 38 88 91 90 -141 136 0 64 -80 -172 54 90 175 -12 -52 -76 112 -105 141 .. 174 I56 31 1 212 NODAL PLANE 2 STRIKE DIP RAKE 307 238 354 212 012 204 278 116 071 159 315 120 304 312 317 108 051 18 1 20 55 55 90 37 30 83 63 7 86 70 66 49 75 54 96 0 90 -30 43 173 130 -108 - 30 112 90 32 -97 -110 -103 84 -79 Downloaded from http://gji.oxfordjournals.org/ by guest on October 21, 2016