INTRODUCTION

        On 3 February 2002, at 09 h 11 min 28 sec GMT the small townships of Sultandağı, Çay and Bolvadin southeast of Afyon in southwest Turkey were shaken by a destructive earthquake (Fig.1) which was strongly felt in the surrounding cities such as Konya, Ankara, Eskişehir, Kütahya and Isparta. During the earthquake 46 lives were lost, 318 persons were wounded and 622 buildings were moderate to heavily damaged. The township Çay was recorded to experience the heaviest damaged, followed by Sultandağı and Bolvadin. In rural areas the intensive damage was recorded in the villages surrounding the the townships Çay and Sultandağı.

         The earthquake is a result of the reactivity of Sultandağı fault which was shown on the Active Fault Map of Turkey prepared and published by General Directorate of Mineral Research and Exploration (MTA) of Turkey in 1992. The former earthquake (Ms=5.8) in the area was recorded on the western termination of this fault zone, in Çobanlar area, on 15 December 2000.

         A group of researchers of the Natural Disaster Research Unit of Geological Research Department of MTA passing to the earthquake area on 4 February 2002 upon the direction of the Director General and has prepared the following preliminary report after a field work of eight days. The report aims to give general information on the earthquake itself and the surface ruptures related. During the field work some difficulties were faced in systematic data collection due to the unfavourable conditions resulted from diurnal freezing and thawing, and snow cover on the ground in regions where the surface ruptures observed as micro cracks. Further research is being carried out on the detailed geometry of faulting and on the setting of the earthquake in the regional active tectonics.

TECTONIC SETTING AND THE SULTANDAĞI FAULT

        Sultandağı earthquake occurred on the northern end of the Isparta bend which holds a significant place in the neotectonic structure of Turkey hampering the westerly escape of Anatolian block. Isparta bend is bounded by the Sultandağı fault in its north-northeastern front. In the Active Fault Map of Turkey prepared and published by MTA in 1992 the Sultandağı fault was described as a reverse active fault. The Afyon-Akşehir basin that was shaped relevant to the movement of the Sultandağı fault is the easternmost section of the graben structures that developed related to the Aegean extensional tectonic regime (Koçyiğit et al., 2000). The Sultandağı fault traced between the Sultandağı rise and the graben basins to the east and north of it has an approximate length of 100 km on the Active Fault Map of Turkey (Fig.2). When the Afyon graben related to the fault is included, the length of the fault reaches up to 150 km. The Sultandağı fault is a low-angle normal fault. The main fault lies in a zone of tens of meters to a few kilometers. Morphotectonic data suggests that the vertical displacement exceeds 1000 m in its northern end around Çay. The 3 February 2002 earthquake occurred on the northwestern end of the fault. During the surface rupture mapping indications on the Holocene activity of the fault in some locations were observed. Fault scarps and deformations in Holocene sediments related to the previous earthquakes are obvious in the area.

         The grabens situated in Sultandağı region form a significant seismogenic zone in Turkey. Intensive microseismic activity and three earthquakes M>5 were recorded in the area, namely, 1921 Argıthanı (M=5.4), 1946 Argıthanı (M=5.5) and 15 December 2000 Çobanlar (Mw=6.0) earthquakes (Ambraseys, 1988, Taymaz and Tan, 2001). The first two of these occurred on the southern termination of the fault, whereas the latest event took place on the northern termination. It can be said that, from historical point of view, 3 February 2002 Sultandağı earthquake occurred on the Çay segment, filling a seismic gap.

SEISMOLOGIC DATA

        The main shock of Sultandağı earthquake was recorded as Ms=6.5 (USGS) on 03 February 2002 at 09 h 11 min 28 sec GMT followed by intensive aftershocks (Fig.3). Two hours after the main shock another event of Ms=5.6 (USGS) occurred and collapsed the buildings damaged during the main shock. The data on the main shock and the second event collected from various sources were given at Tables 1 and 2, respectively. Epicentral location for the main shock is north of the township of Sultandağı according to USGS, Kandilli Observatory and Swiss Seismological Service (ETHZ). Harvard University Seismology Group locates the epicenter of the main shock north of township of Çay. The epicentral locations for the second event are at the western continuation of the Sultandağı fault. Fault plane solutions for both events reflect normal faulting (Fig.2 ). The epicentral location for the second event is very close to that of 15 December 2000 Çobanlar earthquake which also is a normal fault according to fault plane solutions.

Table 1. Seismologic data of the Sultandağı earthquake from various sources (KOERI: Boğaziçi University Kandilli Observatory; DAD: Gen. Dir. of Disaster Affairs Earthquake Research Dept.; USGS: United States Geological Survey; HARVARD: Harvard University Seismology Group, ETHZ: Swiss Seismological Service).

Table 2. Seismologic data of the large aftershock from various sources (KOERI: Boğaziçi University Kandilli Observatory; DAD: Gen. Dir. of Disaster Affairs Earthquake Research Dept.; USGS: United States Geological Survey; HARVARD: Harvard University Seismology Group, ETHZ: Swiss Seismological Service).

SURFACE RUPTURES

         The 3 February 2002 earthquake is a result of the movement of the Sultandağı fault and along which surface ruptures were formed. Between the Maltepe village and Eber creek the ruptures were observed along 21 km (Fig.4) , (Fig.5). The general strike of the surface ruptures is N800W which is conformable with that of Sultandağı fault in Çay area. The ruptures are continuous between Çay township and Maltepe village. From fault geometry point of view the rupture can be divided as Maltepe and Çay segments. The general strikes of Çay and Maltepe segments are E-W and N750W, respectively. In the west the surface ruptures fade out in the swampy alluvial plain whereas in the east they spread along two main directions varying between tens of meters and hundreds of meters. Around and east of Çay surface ruptures are observed in metamorphic basement rocks whereas they cut the Pleistocene and Holocene sediments in Maltepe region, in the plain

         The maximum perpendicular to strike displacement on the rupture measured in the first two days is 25 cm. Extensions generally were observed. The measurements after eight days of the main shock have shown that the values of vertical displacement and separation were increasing on micro ruptures.

         The general strike of the ruptures and the direction of extension are conformable with the fault plane solutions in Çay area. Some right and left lateral offsets were also observed due to changes in the strike. Left lateral strike-slip component is dominant in the sections at the easternmost termination. In the appendix photographs of the rupture from different locations were given.

         In a trench excavated in the slope debris to repair the water pipes damaged during the earthquake near southwest of the Çay township the surface ruptures were observed to follow the fault planes of past earthquakes. Similarly, in the areas where the ruptures display continuity, the fresh fault scarps indicate that the surface ruptures are conformable with the past breaks.

LIQUEFACTION

        Despite the presence of large plains on the hanging wall of the fault, only in some limited areas liquefaction process was observed after the Sultandağı earthquake. The most significant liquefaction examples were found next to Sanayi Sitesi (Industrial Site) which stands on the recent river bed of the Çay creek. 400-500 m northeast of this location we observed some liquefaction cones. In the area many structures of differing size resembling liquefaction cones made up of sand and silt and mostly pebbles were observed. Some of these cones, especially that of containing silt are remarkable with the mud flow structures and are very similar to liquefaction cones. Close examinations, however, have shown that the cones are molehills and the mudflow structures have formed due to sudden snow melt related to variable meteorologic conditions.

LATERAL SPREADING AND MASS MOVEMENTS

        During our research, the shores of Lake Eber and lake Akşehir and Akarçay channel were evaluated as lateral spreading areas. Along the Akarçay channel, Between Kadiköy and Lake Eber small scale landslides rather than lateral spreading were observed. North of Uyanık village, 250-300 m away from the southern shore of Lake Eber, a small scale lateral spreading was observed in the old lake boundary.

        No detailed investigation was done on the mass movements -and their regional distribution - triggered and developed during the earthquake. In some areas, however, rock falls and debris flow are quite obvious; north of Lake Akşehir rock falls at different localities from limestone scarps are remarkable. Along the steep slopes of the valleys south of Sultandağı and Çay rock falls as well as debris flows were observed.

CHANGES IN UNDERGROUND WATER AND GEOTHERMAL SPRINGS

        After the earthquake, north of the surface ruptures, on the hanging wall of the fault water level changes were observed in the wells and springs. The water level in the wells on the plain between Sultandağı and Çobanlar reported to decrease after 15 December 2000 Çobanlar earthquake has contrarily increased and the water gushed after the 03.02.2002 Sultandağı earthquake.

        In Heybeli Kızılkilise geothermal area situated on the west end of the surface ruptures, a hot spring from which the water had been pumped before the earthquake gushed after the earthquake. The water discharge was recorded to increase in the other wells in the area. Three days after the earthquake an extension fissure of 35 m in length and lying in N300E direction with a separation of 0.5-1 cm was observed to develop on the traventine ground in the area. Along the fissure many hot water seeps were seen. The temperatures of the hot spas in this area and in Afyon region were not measured to rise (pers. comm. MTA Energy Resources Department Research Group). However, in Heybeli Kızılkilise geothermal area increase in the gas seep was recorded.

GEOLOGICAL ASPECTS OF DISTRIBUTION OF DAMAGE

        The interaction between the damage and ground conditions is beyond the scope of this research. Some evaluations, however, can be done on the subject. Sultandağı earthquake, when compared in magnitude to the other earthquakes of the same nature in the last 50 years in western Turkey, seems to be slight to moderate in casualties and damage. The 01 October 1995 Dinar earthquake (M=6.0), for example, was more destructive than 03 February 2002 Sultandağı earthquake comparing to its magnitude. This contrast practically can be explained by the quality, type and density of housing. The MTA report on the Dinar earthquake states that the highest damage was recorded on the sandy, swamp areas where the ground water is higher. We have concluded that for Sultandağı earthquake the ground conditions were not much effective on the damage; the structural weakness of the houses and the buildings has given way to collapses.
        Although the surface rupture crosses the township of Çay, no heavy damage due to surface rupturing was recorded. The heaviest damage in the earthquake area is the Çay Industrial Site which stood on the reclamation area of the Çay creek. The studies on an excavation wall next to the site indicate that the area is made up of unconsolidated sand and pebbles of the old river bed.
        On the eastern termination of the surface ruptures where they spread in a zone, it is remarkable that the damage in rural settlements the damage is higher than the other areas. Of these, the heavy damage in Eber village is due to construction type and material; the collapsed houses standing for the last 40-50 years were erected by sun-dried bricks. It is reported that in Yakasenek and Deresenek villages the sand used in construction of the reinforced concrete buildings have been supplied from river beds and alluvial fans in the area where the mud content is too high to use in constructions.

PRELIMINARY RESULTS

        The 03 February 2002 Sultandağı earthquake occurred on the Sultandağı active fault, which was shown on the Active Fault Map of Turkey, prepared and published by MTA (1992). A surface rupture of 21 km reflecting normal faulting features with a maximum displacement of 25 cm was observed. It is highly probable that the earthquake (M= 5.6), which occurred two hours after the main shock, might be a triggered event.

REFERENCES

Ambraseys, N.N., 1988, Engineering seismology. Earthquake Engineering and Structural Dynamics, 17, 1-105.
Koçyiğit, A., Ünay, E. and Saraç, G., 2000, Episodic graben formation and extensional neotectonic regime in west Central Anatolia and the Isparta Angle: a case study in the Akşehir-Afyon Graben, Turkey. Tectonics and Magmatism in Turkey and the Surrounding Area, (Eds.) E. Bozkurt, J.A. Winchester and J.D.A. Piper, Spec. Publ., 173, 405-421.

PHOTOGRAPHS