Active faulting at Delphi, Greece: Seismotectonic remarks and a hypothesis for the geologic environment of a myth Luigi Piccardi* Consiglio Nazionale delle Ricerche, Centro di Studio di Geologia dell’Appennino e delle Catene Perimediterranee, Firenze, Italy
ABSTRACT Historical data are fundamental to the understanding of the seismic history of an area. At the same time, knowledge of the active tectonic processes allows us to understand how earthquakes have been perceived by past cultures. Delphi is one of the principal archaeological sites of Greece, the main oracle of Apollo. It was by far the most venerated oracle of the Greek ancient world. According to tradition, the mantic proprieties of the oracle were obtained from an open chasm in the earth. Delphi is directly above one of the main antithetic active faults of the Gulf of Corinth Rift, which bounds Mount Parnassus to the south. The geometry of the fault and slipparallel lineations on the main fault plane indicate normal movement, with minor right-lateral slip component. Combining tectonic data, archaeological evidence, historical sources, and a reexamination of myths, it appears that the Helice earthquake of 373 B.C. ruptured not only the master fault of the Gulf of Corinth Rift at Helice, but also the antithetic fault at Delphi, similarly to the Corinth earthquake of 1981. Moreover, the presence of an active fault directly below the temples of the oldest sanctuary suggests that the mythological oracular chasm might well have been an ancient tectonic surface rupture. Keywords: seismotectonics, earthquakes, surface faulting, Gulf of Corinth, Delphi. INTRODUCTION The Gulf of Corinth Rift, one of the most prominent tectonic features in the Aegean Sea region, is an asymmetric crustal half-graben trending roughly east-west (Fig. 1). The master fault system, dipping northward, is at its southern margin, while minor antithetic faults affect the downwardflexed, northern hanging wall (Jackson et al., 1982; Taymaz et al., 1991; Armijo et al., 1996). Fast active extension in the region, approximately north-south oriented (McKenzie, 1978; Jackson et al., 1982), is expressed by high seismicity. Maximum earthquake magnitudes exceed 6 (Mariolakos et al., 1989; Ambraseys and Jackson, 1990; Ambraseys, 1996). The occurrence of coseismic surface faulting is a well-known phenomenon in this area, as shown in at least three cases (Fig. 1). In the 1981 Corinth seismic sequence, normal coseismic surface faulting, coherent with fault-plane solutions, was observed on two main conjugate faults at the eastern end of the Corinth graben (Jackson et al., 1982). The shocks of February 24 and 25, 1981 (Ms = 6.7 and Ms = 6.4, respectively) were associated with fault motion on the south side of the gulf (Skinos fault). With the shock of March 4, 1981 (Ms = 6.4), breaks opened along an antithetic fault on the north side of the graben (Kaparelli fault). In 1861, the Egion earthquake produced a rupture of ~13 km along the Helice active fault (Schmidt, 1881; Mouyaris et al., 1992). Similarly, in the case of the Helice earthquake of 373 B.C., Mouyaris et al. (1992) proposed surface normal faulting to explain the “chasm of the Earth,” which classical authors claimed to have swallowed the town of Bura. Integrating tectonic, archaeological, and historical data, this paper examines the hypothesis that the earthquake of 373 B.C. ruptured not only the Helice fault on the southern border of the graben, but also the corresponding antithetic fault at Delphi (Fig. 1). If this is the case, the seismotectonic mechanism of that earthquake would appear to be the same as that of the 1981 Corinth earthquake. Furthermore, we used the interdisciplinary approach to investigate the geologic background of the Delphic myth, as has been done in similar cases (e.g., Bentor, 1989; Guidoboni, 1989; Nur, 1991; Armijo et al., 1991; Nur and Ron, 1996; Piccardi, 1998). Delphi was the seat of the most famous *E-mail:
[email protected]. Geology; July 2000; v. 28; no. 7; p. 651–654; 4 figures.
Greek oracle, the main sanctuary of Apollo. The priestess is said to have uttered her oracles in a state of frenzy induced by exhalations, which arose from a natural open chasm in the earth, directly on top of which the main temple was erected (e.g., Diodorus, 1952, 16.26.1-6). Until now, this cleft had not been identified, and most scholars have considered it to be merely a mythological invention (e.g., Parke and Wormel, 1956; Fontenrose, 1981). TECTONICS OF THE AREA The Gulf of Corinth Rift is characterized by particularly high rates of deformation, which are among the highest in the Mediterranean area. The current phase of extension started ca. 1 Ma (Jackson et al., 1982; Lyon-Caen et al., 1988; Ori, 1989; Mouyaris et al., 1992; Armijo et al., 1996), and has an average rate of about 10 mm · yr –1 (Ambraseys and Jackson, 1990; Billiris et al., 1991). On the basis of geologic data and uplifted topography, vertical slip rates of the master fault system appear to range between 6–7 and 11 mm· yr –1 at Xilocastro (Armijo et al., 1996) and between 3 and 8 mm· yr –1 at Helice (Mouyaris et al., 1992; Armijo et al., 1996). Paleoseismological trenching indicated a slip rate between 0.7 and 2.5 mm · yr –1 for the Skinos fault alone (Collier et al., 1998). The Delphi fault bounds the Mount Parnassus massif to the south (Figs. 1 and 2) and it is the most prominent of the active antithetic structures of the Gulf of Corinth Rift (Péchoux, 1977; Bousquet et al., 1977; Armijo et al., 1996; Ambraseys, 1996). It is an east-west–trending, normal fault, dipping to the south, with minor right-lateral component of movement. Geologic and geomorphic features indicate that the fault is active (Figs. 2 and 3A). Assuming the 500–700-m-high mountain front to have originated by slip on the Delphi fault in the last extensional phase (ca. 1 Ma), the average slip rate would range between 0.5 and 0.7 mm/yr. For the most part, the main fault plane separates Mesozoic limestone and Paleogene flysch of the bedrock from Quaternary slope debris. At site 1 (Fig. 2) the fault plane has an average strike of 260°N, dips 60°–65°, and is lineated by corrugations (wavelength to 2 m) and striations with a pitch of about 70°W. Wurmian slope debris and Holocene debris cones are faulted as well (Fig. 3B). Gas emissions described by Fontenrose (1981), and travertine deposits that in places cover archaeological relics, indicate hydrothermal activity in the area. Ancient Delphi was located right above this fault (Ap, K, and At in Figs. 2 and 3A). 651
22°20′E
Elev. (ft)
22°E
38°40′N
22°40′E
Major active faults Thermopylai Minor active faults Focal mechanisms Epicenter of 373 B.C.
Bathymetry (m) Tsunamis
23°E
Coseismic ruptures 1981 1861 373 B.C. 373 B.C. at Delphi (this paper)
Greece Turkey
Fig.1 He ll e
Mt. Parnassus
c
ni
Figure 1. Eastern part of Gulf of Corinth. Topography (in feet) is from U.S. Defense Mapping Agency (1992). Bathymetry, in meters, is from Pessioratis et al. (1986). Active faults are modified from Mariolakos et al. (1989) and Armijo et al. (1996). Fault-plane solutions are from Taymaz et al. (1991). 1981 Corinth earthquake ruptures are from Jackson et al. (1982). 1861 Egion earthquake rupture is from Schmidt (1881) and Mouyaris et al. (1992). 373 B.C. Helice earthquake rupture on Helice fault is from Mouyaris et al. (1992). Epicenter of 373 B.C. earthquake is from Ambraseys (1996). Cross section was redrawn from Armijo et al. (1996).
600
11000 7000 3000 0
A
Aegean Sea Tr e nc
h
Delphi Fig. 2
5.03.81
11.02.84
38°20′N
Egion
373 B.C.
400
Kaparelli
Helice 600 Bura
Gulf
of C orinth
800
25.02.81
24.02.81
4.03.81
Xylocastro
ADelphi
Skinos
B
Gulf of Corinth
38°N
0
Vertical Exaggeration = 2 5 40
22°28′E 38°30′N
22°29′E
22°30′E
0
20
22°31′E
1164
1225 1205
K Ap
Delphi
22°33′E
Holocene alluvial deposits Quaternary talus and slope fan debris 1306 Quaternary compact talus debris Paleogene flysch and thin-bedded limestones Mesozoic limestones
1
At
Fig. 4A inset
Fig.3B Chrysa
Pleistos
R.
N
277
1
38°28′N
Fig.3A
1
Major active fault, with slip vector
Figure 3. A: General view of site. Although morphology is complicated at places by strata orientation, mountain front is faceted with triangular and trapezoidal spurs (f), separated by distinct hanging valley (hv1). Facets join along their bases, which are marked by steeper morphological scarps. Hanging valley (hv2) is also visible at higher elevation. B: Line drawing of view toward east of shrine of Athena.
Figure 2. Schematic geologic map of area of Delphi (modified from Zachos, 1964). Slipparallel lineations, measured on main fault plane (site 1), indicate north-northeast–southsouthwest direction of extension, reflecting actual general direction of extension of area. Right-lateral component of slip is suggested also by geometry of northwest-trending fault bend (between K and At).
779
0
Minor active fault
1
2
km
3
4
800
Dip of strata
View of photographs
5
K : Kastalia 760 Spring At : Shrine of Athena Ap : Shrine of Apollo
Mt. Parnassus 2457 m
A
hv2
W f
N
B
E
site1
f hv1 f
Shrine of Apollo
Modern DELPHI
652
Corinth
B
5 km
22°32′E
1177
38°29′N
20
Ap
hv1
Altar
Temple
f
At
K At
Kastalia Spring
Fig. 3B
Shrine of Athena
S Debris cones Paleosurface Quaternary Bedrock
GEOLOGY, July 2000
Intrafault hanging-wall collapse (Stewart and Hancock, 1988) creates a stepped morphology. Movement is distributed at the surface on a few parallel subsidiary faults, within a fault zone several tens of meters wide (Péchoux, 1977). This is particularly visible to the east of Kastalia Spring (Fig. 3B), where subsidiary faults offset the paleosurface of the Quaternary slope deposits. Slip on the most external of these faults during historical times may be inferred from archaeological evidence. In the shrine of Athena this fault cuts and displaces the main temple and altar (end of sixth century B.C., Fig. 4A). Offset of about 30 cm, still evident in the main altar’s west wall (Fig. 4B), caused the break, and tilt, of the basement of the temple. EARTHQUAKE OF 373 B.C. The most famous historical earthquake in this sector of the Corinthian Gulf has been the one of 373 B.C. (Fig. 1). The earthquake struck at night, destroying two cities on the southern border of the gulf. Bura is said to have “disappeared in a chasm of the Earth,” while Helice was “wiped out by a wave of the sea” the following morning, and remained submerged (e.g.,
K Kastalia
Ap
Shrine of Apollo Modern Delphi
N
Spring
A
N
Fig.4A Shrine of Athena At
0
500 m
0 10 20 30 40 50 m
IIa
IIIa
Ia
T
Ib IIb
Fig.4B
sacred pre
cinct
Temple Ia Main Altar Ib
Temple IIa Main Altar IIb
(ca. 650-630 B.C.)
(ca. end of sixth T: Tholos century B.C.) (ca. 380-370 B.C.)
Temple IIIa (ca. 365-360 B.C.) 510
N
Original sacred area Extension after 373 B.C. Fault trace View of photograph
B
IIb Ib S Figure 4. Setting of archaeological site (modified from Bommelaer, 1991).Temple of Apollo that remains today (Ap) dates to fourth century B.C. Previous temple, built on same location in 514–506 B.C., was destroyed by earthquake of 373 B.C. Its seventh century B.C. predecessor burned in 547 B.C. Three earlier temples of Apollo are described in legends, but their locations remain unknown. A: Shrine of Athena Pronaia. Only last temple, IIIa, is known to have been dedicated to Athena. B: Main altar IIb. Offset of west wall is visible. In background, morphologic scarp is aligned with this rupture. GEOLOGY, July 2000
Strabo, 1917, 1.3.18; Diodorus, 1952, 15.1-4). Mouyaris et al. (1992) proposed that rupture of the Helice fault, with associated surface faulting, is the most probable source of this earthquake. That earthquake caused massive destruction at Delphi too, although far away from the main-shock epicenter. From archaeological data, Bousquet and Pechoux (1977) attributed an intensity of IX MCS to this shock at Delphi. The site was reconstructed after the event. Fabric and style of the new buildings are much poorer in quality compared to the previous ones, due to the lack of economic resources (Parke and Wormel, 1956; Bommelaer, 1991). It is worth noting that, in the reconstruction, the main temple of Athena was moved from its original position (from IIa to IIIa in Fig. 4A), and rebuilt smaller than before. Plutarch (1936) reported some old oral traditions that spoke of permanent traces left somewhere in the sanctuary after this earthquake, and that were still present in his time (early second century). Apparently, the unrecoverable and/or too-expensive-to-repair damages to the temple are likely to have been the reason for this change of location. This suggests that faulting of the basement of the altar and temple occurred during the earthquake of 373 B.C. ORACULAR CHASM Since the origins of Delphi, earthquakes are one of the leading themes of its mythology (e.g., Parke and Wormel, 1956; Polimenakos, 1996). According to tradition, Delphi was originally the oracle of Ge, Mother Earth, and Poseidon, the Earth-shaker. Apollo is said to have maintained the original location of the oracle, after he had acquired it by force, having slain the dragon-snake guardian (Delphine or Python). The first description of the slaying of the serpent (Homeric Hymn to Apollo, ca. sixth century B.C. [see Homer, 1936]), recalls an earthquake scenario, albeit clothed in mythological images: the monster “lay drawing great gasps for breath, and rolling about that place. Awful noise swelled up unspeakable, as she writhed continually this way and that amid the wood.” From the rotting of the snake, the site’s alternative name, Pytho (= rotten), is derived. The existence of the sacred chasm (“stomion”) has been a subject of debate since Hellenistic time. It was first mentioned in Aeschylus’s (1957) Oresteia (458 B.C.), and was always indicated as belonging to Ge. The primitive sanctuary of fourteenth century B.C. occupied the same position as the shrine of Athena (Bommelaer, 1991). Therefore, it is here that the oracular chasm, later kept by Apollo, must have been originally located. This exclusively feminine sanctuary might well have been that forbidden sanctuary in which Diodorus (1952, 16.26) described the sacred cleft. As previously shown, geologic and seismological data indicate an active fault cutting across the shrine of Athena (Fig. 4). This fault, although probably including gravitational-sliding components, is likely to have been the seat of coseismic surface ruptures in the past. Such a deep fissure in the ground, with typically associated gas emissions, could correspond to the mythological chasm. CONCLUDING REMARKS The sanctuary of Delphi, in existence at least since 1400 B.C., was closed in A.D. 381, after uninterrupted worship for almost 2000 years. Given the location of the archaeological site above the Delphi active fault, its long history, and the frequency of earthquakes, it is not surprising to find here historical and archaeological evidence of surface faulting. Telluric manifestations on this active fault seem to be the reason for the veneration of the place. The local tectonic framework might, in fact, explain most of the definite correspondences existing between the geologic features, the archaeological evidence, and the content of local myths. The presence of the active subsidiary fault in the shrine of Athena, the original pre-Apolline sanctuary, suggests that the mythological chasm of goddess Earth could have been an ancient coseismic surface rupture. A similar chasm can remain open for several decades, especially when protected inside a restricted access area; however, it will eventually close. Likewise, the chasm could have existed long enough to give rise to a well-established 653
myth, only to close and be forgotten centuries later. The strongly asymmetric organization of this shrine, the very unusual orientation of the main temples and lateral positioning of the main altar (Bommelaer, 1991; Fig. 4), supports evidence of an intentional placement of the buildings directly above the fault trace. This temple and altar are therefore perfect archaeological seismological markers. They appear to have been broken by slip on the fault, most probably in the earthquake of 373 B.C. The Helice and Delphi faults relate tectonically to each other in a similar way, although further apart, like the Skinos and Kaparelli faults (Fig. 1). A seismotectonic mechanism implying ruptures on both synthetic and antithetic faults of the graben, as happened in the 1981 Corinth seismic sequence, could therefore be suggested for the 373 B.C. earthquake as well. ACKNOWLEDGMENTS I am grateful to E. Abbate, R. Armijo, M. Boccaletti, G. Moratti, C. Poccianti, and F. Sani, for critically reading the manuscript, discussions, and encouragement. Constructive reviews by A. Nur and S. Schweig greatly improved the paper. I also owe thanks to F. Piccardi and J. Arkell for revision of the text in English. This is C.N.R.–Centro di Studio di Geologia dell’Appennino e delle Catene Perimediterranee contribution no. 333. REFERENCES CITED Aeschylus, 1957, Oresteia: Loeb Classical Library, Volume 146 (eleventh edition): Cambridge, Massachusetts, Harvard University Press, 618 p. Ambraseys, N., 1996, Material for the investigation of the seismicity of central Greece, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Exeter, UK, Fitch Laboratory Occasional Paper 7, p. 23–36. Ambraseys, N., and Jackson, J., 1990, Seismicity and associated strain of central Greece between 1890 and 1988: Geophysical Journal International, v. 101, p. 663–708. Armijo, R., Lyon-Caen, H., and Papanastassiou, D., 1991, A possible normal-fault rupture for the 464 B.C. Sparta earthquake: Nature, v. 351, p. 137–139. Armijo, R., Meyer, B., King, G.C.P., Rigo, A., and Papanastassiou, D., 1996, Quaternary evolution of the Corinth Rift and its implications for late Cenozoic evolution of the Aegean: Geophysical Journal International, v. 126, p. 11–53. Bentor, Y.K., 1989, Geological events in the Bible: Terra Nova, v. 1, p. 326–338. Billiris, H., Paradissis, D., Veis, G., England, P., Featherstone, W., Parsons, B., Cross, P., Rands, P., Rayson, M., Sellers, P., Ashkenazi, V., Davison, M., Jackson, J., and Ambraseys, N., 1991, Geodetic determination of tectonic deformation in central Greece from 1900 to 1988: Nature, v. 350, p. 124–129. Bommelaer, J.F., 1991, Guide de Delphes: Paris, Ecole Francaise d’Athènes, De Boccard Edition, 541 p. Bousquet, B., and Péchoux, P.Y., 1977, La seismicité du Bassin égéen pendant l’Antiquité. Méthodologie et premiers résultats: Bulletin de la Société Géologique de France, v. 19, p. 679–684. Bousquet, B., Dufaure, J.J., and Péchoux, P.Y., 1977, Le rôle de la géomorphologie dans l’évaluation des déformations néotectoniques en Grèce: Bulletin de la Société Géologique de France, v. 19, p. 685–693. Collier, R.E.L., Pantosti, D., D’Addenzio, G., De Martini, P.M., Masana, E., and Sakellariou, D., 1998, Paleoseismicity of the 1981 Corinth earthquake fault: Seismic contribution to extensional strain in central Greece and implications for seismic hazard: Journal of Geophysical Research, v. 103, no. B12, p. 30,001–30,019. Diodorus, 1952, Siculus: Historical Library, XV–XVI: Loeb Classical Library, Volume 389 (fifth edition): Cambridge, Massachusetts, Harvard University Press, 438 p. Fontenrose, J., 1981, The Delfic Oracle, its responses and operations: Berkeley, California, University of California Press, 476 p. Guidoboni, E., ed., 1989, I terremoti prima del Mille in Italia e nell’area Mediterranea: Bologna, Società Geofisica Ambiente, 765 p. Homer, 1936, Hymn to Apollo: Loeb Classical Library, Volume 57 (seventeenth edition): Cambridge, Massachusetts, Harvard University Press, 706 p.
654
Jackson, J.A., Gaignepain, J., Houseman, G., King, G.C.P., Papadimitriou, P., Soufleris, C., and Virieux, J., 1982, Seismicity, normal faulting and the geomorphological development of the Gulf of Corinth (Greece): The Corinth earthquakes of February and March 1981: Earth and Planetary Science Letters, v. 57, p. 377–397. Lyon-Caen, H., Armijo, R., Drakopolous, J., Baskoutass, J., Delibassis, N., Gaulen, R., Kouskouna, V., Latoussakis, J., Makropoulos, K., Papadimitriou, P., Papanastassiou, D., and Pedotti, G., 1988, The Kalamata (South Peloponnesus) earthquake: Detailed study of a normal fault, evidences for east-west extension in the Hellenic Arc: Journal of Geophysical Research, v. 93, no. B12, p. 14,967–15,000. Mariolakos, E., Bornovas, J., and Mouyaris, N., eds., 1989, Seismotectonic map of Greece, with seismological data: Athens, Institute of Geology and Mineral Exploration, scale 1:500 000. McKenzie, D.P., 1978, Active tectonics of the Alpine-Himalayan belt: The Aegean Sea and surrounding regions: Royal Astronomical Society Geophysical Journal, v. 55, p. 217–254. Mouyaris, N., Papastamatiou, D., and Vita-Finzi, C., 1992, The Helice fault?: Terra Nova, v. 4, p. 124–129. Nur, A., 1991, Earthquakes in the Bible: New Scientist, v. 1776, p. 45–48. Nur, A., and Ron, H., 1996, And the walls came tumbling down: Earthquake history in the Holy Land, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Exeter, UK, Fitch Laboratory Occasional Paper 7, p. 75–85. Ori, G.G., 1989, Geologic history of the extensional basin of the Gulf of Corinth (?Miocene-Pleistocene), Greece: Geology, v. 17, p. 918–921. Parke, H.W., and Wormel, D.E.W., 1956, The Delphic Oracle, the history, Volume 1: Oxford, UK, Blackwell, 436 p. Péchoux, P.Y., 1977, Nouvelles remarques sur les versants Quaternaires du secteur de Delphes: Revue de Géographie Physique et de Géologie Dynamique, v. 19, p. 83–92. Pessioratis, C., Mitropoulos, D., and Angelopoulos, I., 1986, Marine geological research at the E Korinthiakos Gulf: Special Issue of Geology and Geophysical Research: Athens, Institute of Geology and Mineral Exploration, p. 381–401. Piccardi, L., 1998, Cinematica attuale, comportamento sismico e sismologia storica della faglia di Monte Sant’Angelo (Gargano, Italia): La possibile rottura superficiale del “leggendario” terremoto del 493 d.C.: Geografia Fisica e Dinamica Quaternaria, v. 21, p. 155–166. Plutarch, 1936, The oracles at Delphi no longer given in verse—The obsolence of oracles: Loeb Classical Library, Volume 306: Cambridge, Massachusetts, Harvard University Press, 528 p. Polimenakos, L.C., 1996, Thoughts on the perception of the earthquake in Greek antiquity, in Stiros, S., and Jones, R.E., eds., Archaeoseismology: Exeter, UK, Fitch Laboratory Occasional Paper 7, p. 253–257. Schmidt, J.F.J., 1881, Studien über vulcans und erdbeben: Leipzig, Carl Scholtz Ed. Stewart, I.S., and Hancock, P.L., 1988, Normal fault zone evolution and fault scarp degradation in the Aegean region: Basin Research, v. 1, p. 139–153. Strabo, 1917, Geography, I: Loeb Classical Library, Volume 49 (seventh edition): Cambridge, Massachusetts, Harvard University Press, 547 p. Taymaz, T., Jackson, J., and McKenzie, D., 1991, Active tectonics of the north and central Aegean Sea: Geophysical Journal International, v. 106, p. 433–490. U.S. Defense Mapping Agency, 1992, Digital chart of the world: Fairfax, Virginia, scale 1:1 000 000. Zachos, K., ed., 1964, Geological map of Greece, Delphi: Athens, Institute of Geology and Subsurface Research, scale 1:50 000. Manuscript received November 23, 1999 Revised manuscript received March 30, 2000 Manuscript accepted April 20, 2000
Printed in USA
GEOLOGY, July 2000
