ARTICLE IN PRESS Quaternary Geochronology 2 (2007) 284–289 www.elsevier.com/locate/quageo Research paper Luminescence dating of the last earthquake of the Sabzevar thrust fault, NE Iran Morteza Fattahia,b,, Richard T. Walkerc a The Institute of Geophysics, University of Tehran, Kargar Shomali, Tehran, Iran OLRG, School of Geography, University of Oxford, South Parks Road, Oxford OX1 3QY, UK c COMET, Department of Earth Sciences, University of Oxford, Parks Road, Oxford OX1 3PR, UK b Received 24 May 2006; accepted 7 June 2006 Available online 25 July 2006 Abstract Iran is one of the world’s most tectonically active regions, yet dating past earthquakes for neotectonic studies has been limited. One of the main reasons for this is that organic material suitable for radiocarbon dating of deformed sediments is rare. We investigate the use of infrared stimulated luminescence (IRSL) from coarse-grained feldspars to date colluvial deposits associated with the Sabzevar thrust fault in northeastern Iran. The single-aliquot regenerative (SAR) dose measurement procedure was used for this study. The current study investigates monitoring and correcting for sensitivity changes, recovering a known laboratory dose and equivalent dose estimation using three SAR IRSL methods. It is shown that SAR has recovered a given laboratory dose using a range of preheat temperatures but De determination of natural samples requires its own preheat plateaus for two of these SAR methods. The SAR IRSL method provided an age of 1.770.3 ka for colluvium, predating the last earthquake event on the Sabzevar fault. This result suggests that this fault is likely to be responsible for an earthquake that destroyed Sabzevar city in AD 1052. r 2006 Elsevier Ltd. All rights reserved. Keywords: Luminescence dating; Fault slip-rate; Iran; Earthquake; Seismic hazard 1. Introduction Iran is one of the most seismically active regions along the Alpine-Himalayan belts with numerous destructive earthquakes recorded both historically and instrumentally. For example, an earthquake on the 26th December 2003, with a moment magnitude (Mw) of 6.5, resulted in the loss of over 30,000 lives and almost totally reduced the ancient city of Bam and surrounding villages to ruins (e.g., Talebian et al., 2004). The city of Sabzevar in NE Iran has been relatively free from earthquakes in the modern age, although slight damage was caused by two small earthquakes (with magnitudes of 4.6 and 4.2), on the 12th and 17th December 2004 (from, Institute of Geophysics, Tehran University). Corresponding author. OLRG, School of Geography, University of Oxford, South Parks Road, Oxford OX1 3QY, UK. Tel.: +44 01865 556407; fax: +98 21 8009560, +44 1865 275885. E-mail addresses: morteza.fattahi@ouce.ox.ac.uk, m.fattahi@ut.ac.ir (M. Fattahi). 1871-1014/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.quageo.2006.06.006 However, Sabzevar remains at risk from earthquakes in the future, and historical sources describe how the city was destroyed by a large earthquake in AD 1052 (named the Baihaq earthquake; Ambraseys and Melville, 1982). As the Sabzevar thrust is the major identifiable active fault in the region, and passes very close to the city (Fig. S1), it seems likely that this fault was responsible for the AD 1052 event. However, as surface ruptures were not recorded at the time (Ambraseys and Melville, 1982), this link remains to some extent conjectural. 2. Sampling site and experimental treatment Following each earthquake on a thrust fault in which slip reaches the Earth’s surface, surficial processes modify and gradually degrade the fault scarp, by eroding material from the (uplifted) hanging-wall side of the fault, and depositing this sediment on the (downthrown) footwall side of the fault, forming a wedge of colluvial sediment. An exposure through the Sabzevar fault scarp at 36:13:18N 57:31:33E ARTICLE IN PRESS M. Fattahi, R.T. Walker / Quaternary Geochronology 2 (2007) 284–289 revealed a colluvial wedge, presumably relating to several earthquake events. The uppermost layer of colluvium was clearly cut by faulting (Fig. S2), suggesting that the most recent earthquake post-dates the exposed colluvium. The age of these colluvial sediments therefore provide a valuable constraint on the maximum age of the last faulting event on the Sabzevar fault. The likely interval between large earthquakes on the Sabzevar fault, and hence the local earthquake hazard, can be estimated by combining our results with estimates of the average fault slip-rate (Fattahi et al., 2006). One sample (sample S6) was collected from the uppermost layer of colluvium. The infrared stimulated luminescence (IRSL) signal of this sample should be completely reset if deposition of this post faulting sediment has been sufficiently slow. Previous studies have shown that the rapidly bleached IRSL signal in alkali feldspars (e.g., Hutt et al., 1988) is reset during colluvial depositional processes (e.g., Porat et al., 1996). The modified single-aliquot regenerative (SAR) dose protocol of quartz (Murray and Wintle, 2000) was applied to aliquots of 90–150 mm feldspar, which were prepared by wet sieving, HCL and H2O2 treatment, followed by heavy liquid separation (o2.58 g/cm3). All the experiments reported here were carried out using a Risø automated TL/OSL system (Model TL/OSL-DA-15; fitted with a 90 Sr/90Y beta source delivering 5 Gy min1) equipped with an IR laser diode ðl ¼ 830 nmÞ as stimulation source. The intensity of light incident on the sample was about 400 mW cm2. IRSL was detected using a electron tubes bialkaline PMT. Luminescence was measured through 7 mm Hoya U-340 filters. 3. Luminescence dating/characteristics Murray and Wintle (2000) introduced the SAR dose protocol for quartz which uses the luminescence signal from a test dose administrated after the regeneration dose luminescence measurement to monitor, and then correct for, any sample sensitivity changes during the measurement process. Several workers tried to extend the SAR protocol 285 to coarse-grain feldspar and to recover known laboratory doses and estimate the De value. Some have reported that SAR can successfully recover a known laboratory dose (e.g., Fattahi, 2001; Fattahi and Stokes, 2004; Preusser, 2003; Blair et al., 2005). Others have found an underestimation of the known laboratory dose (e.g., Wallinga et al., 2000a,b). However, these studies have used different preheat, cut heat and stimulation temperature. This work focuses on the testing of assumptions for three different SAR procedures (Table 1). The first procedure employs a cut heat at 220 1C and stimulation temperature at 50 1C (e.g., Wallinga et al., 2000a,b) The second one applies a cut heat at 220 1C and stimulation temperature at 125 1C (e.g., Preusser, 2003). The third method uses equal preheat and cut heat with IR measuring temperature at 150 1C (e.g., Fattahi, 2001; Blair et al., 2005). 3.1. Sensitivity changes and corrections Sensitivity changes were checked by repeated (7 times) cycles in SAR procedure with a repeated fixed regeneration and test dose. The fundamental assumption in SAR protocol is that if a plot of regeneration dose IRSL (Lx) vs. test dose IRSL (Tx) shows a straight line that passes through the origin, the sensitivity-correction procedure has worked properly (Murray and Wintle, 2000). The above procedure was applied to check the validity of a test dose to monitor and correct the sensitivity changes using three different heating methods. The regeneration dose and test dose were 2.0 and 1.2 Gy, respectively. All IRSL measurements were for 100 s. The preheat temperatures were 200, 230, 250, 270 and 290 1C for the three procedures. Three aliquots were used for each measurement. The results of the average of the three aliquots for each measurement are shown for three methods in Fig. 1. For preheat temperature up to 270 1C, in the method in which preheat and cut heat are equal, the linear relationship passes through the origin (Fig. 1c). For two other methods (when the cut heat and preheat temperature are not equal in both time and temperature) there is no linear relationship which passes through the origin and Lx and Tx Table 1 Generalized single-aliquot regenerated sequence and outline of the steps involved in the three different SAR methods Step Treatment 1 Treatment 2 Treatment 3 Ob.a 1 2 3 4 5 6 7 Give dose Pre-heat (210–290 1C) Stimulation (at 50 1C) Give test dose Cut heat (220 1C) Stimulation (at 50 1C) Return to 1 Give dose Pre-heat (210–290 1C) Stimulation (at 125 1C) Give test dose Cut heat (220 1C) Stimulation (at 50 1C) Return to 1 Give dose Pre-heat (210–290 1C) Stimulation (at 150 1C) Give test dose Cut heat ¼ Preheat Stimulation (at 50 1C) Return to 1 – – Lx – – Tx – Note: In step 2, the sample has been heated to the pre-heat temperature using TL and held at that temperature for 10 s. a Observed: Lx and Tx are derived from the initial IRSL signal (5 s) minus a background estimated from the last part of the stimulation curve. Corrected natural signal N ¼ L0 =T 0 ; Corrected regenerated signal Rx ¼ Lx =T x ðx ¼ 125Þ. ARTICLE IN PRESS M. Fattahi, R.T. Walker / Quaternary Geochronology 2 (2007) 284–289 Equivalent dose (Gy) 300 250 200 150 100 50 100 200 300 400 Regen.Dose IRSL, Lx (counts) 500 400 350 300 250 200 150 100 50 0 Equivalent dose (GY) Test Dose IRSL, Tx (counts) 220 240 260 280 300 280 300 Preheat Temperature (°C) (a) 0 4 3.5 3 2.5 2 1.5 1 0.5 0 200 220 900 800 700 600 500 400 300 200 100 0 100 200 300 400 500 Regen.Dose IRSL, Lx (counts) 600 210 230 250 270 290 0 200 (c) 240 260 Preheat Temperature (°C) (b) 0 Test Dose IRSL, Tx (counts) 4 3.5 3 2.5 2 1.5 1 0.5 0 200 0 Equivalent dose (Gy) Test Dose IRSL, Tx (counts) 286 4 3.5 3 2.5 2 1.5 1 0.5 0 200 Dose Recovery 220 240 260 Preheat Temperature (°C) 280 300 Fig. 2. Dose recovery test. (a) The cut heat was fixed at 220 1C and sample temperature was 50 1C. (b) The cut heat was fixed at 220 1C and sample temperature was 125 1C. (c) The cut heat was equal to preheat and sample temperature was 150 1C. The dashed lines are presented to show the dose to be recovered. 400 600 800 1000 1200 Regen.Dose IRSL, Lx (counts) 1400 Fig. 1. Sensitivity correction tests using different preheating temperature shown in the figures. (a) The cut heat was fixed at 220 and sample temperature was 50 1C. (b) The cut heat was fixed at 220 and sample temperature was 125 1C. (c) The cut heat was equal to preheat and sample temperature was 150 1C. The dashed lines are the trend lines. are not correlated after 250 1C preheat temperature (Fig. 1a, b). There is a one-to-one relationship between Lx and Tx up to 250 1C for the second method (IRSL temperature at 125 1C). 3.2. Thermal transfer and dose recovery tests To test thermal transfer of charge into the IRSL trap as a result of preheating (e.g., Rhodes, 2000) the natural aliquots were stimulated twice at room temperature and IRSL was measured for 100 s, with more than 4 h delay between stimulations (to empty the rapidly bleaching trap). No IRSL signal was observed for the second measurement. This suggests that thermal transfer is not a likely source of uncertainty in these aliquots. Dose recovery tests were carried out to provide a method to determine whether the overall effects of sensitivity changes had been properly corrected for. Three aliquots were used for each preheat temperature. After depleting the natural signal, each aliquot was given 2.6 Gy beta doses and this dose was measured using the three above mentioned SAR procedures (Table 1) and results are shown in Fig. 2. Although all three methods have successfully recovered the laboratory dose, the accuracy of the second method is the best. 3.3. De determination and dating The equivalent dose (De) preheat plateau (Fig. 3) was obtained using the three single-aliquot regeneration methods (Table 1). The preheat time used for all measurements (210–290 1C) was for 10 s. IRSL was measured for 100 s for three methods at 50, 125 and 150 1C sample temperature, respectively. Three disks were prepared for each preheat temperature and following measuring the natural dose, a dose–response curve was constructed from five dose points including three regenerative doses (1.5, 2 and 4 Gy), and a zero dose. A replicate measurement of the lowest regenerative dose ARTICLE IN PRESS Equivalent dose (Gy) M. Fattahi, R.T. Walker / Quaternary Geochronology 2 (2007) 284–289 4 3.5 3 2.5 2 1.5 1 0.5 0 200 220 Equivalent dose (Gy) 4 3.5 3 2.5 2 1.5 1 0.5 0 200 Equivalent dose (Gy) 260 280 300 220 240 260 280 300 Preheat Temperature (°C) 4 3.5 3 2.5 2 1.5 1 0.5 0 200 was carried out at the end of each SAR cycle. The net initial IRSL signal (first 5 s—average of 90–100 s) was used for natural, regenerated and test dose measurements. The De was determined by interpolation and the sensitivity was corrected by dividing Lx by Tx. No aliquot produced significant recuperation signals and all produced recycling ratio between 0.90 and 1.10. The dose recovery test suggested the second SAR method as the most accurate method (Fig. 2b). There is also a clear plateau in the second method for De values (mean De ¼ 2.5470.2 Gy) in the preheat temperature range of 230–290 1C. Therefore, we used this mean value for age determination. 3.4. Anomalous fading test (b) (c) 240 Preheat Temperature (°C) (a) 287 Dates 220 240 260 280 300 Preheat Temperature (°C) 1.2 1 0.8 0.6 0.4 0.2 0 Lx/Tx Lx/Tx Fig. 3. Plot of equivalent dose as a function of preheat temperature. (a) The cut heat was fixed at 220 1C and sample temperature was 50 1C. (b) The cut heat was fixed at 220 1C and sample temperature was 125 1C. (c) The cut heat was equal to preheat and sample temperature was 150 1C. The dashed lines are presented to show the accepted value of De (2.5470.08 Gy) for age determination. The scattering of De measured for each preheat temperature is shown by large variability in the error bars. 0 1 2 6 7 8 1.2 1 0.8 0.6 0.4 0.2 0 0 1 2 230°C 3 4 5 Cycle no. 6 7 1 2 3 4 5 Cycle no. 1.2 1 0.8 0.6 0.4 0.2 0 8 0 1 2 3 4 5 Cycle no. 6 7 8 1 2 3 6 7 8 6 7 8 290°C 1.2 1 0.8 0.6 0.4 0.2 0 0 1.2 1 0.8 0.6 0.4 0.2 0 0 270°C Lx/Tx Lx/Tx 3 4 5 Cycle no. Lx/Tx Lx/Tx 200°C 250°C A fading test was performed by repeated (7 times) cycles of the SAR procedure with a fixed regeneration (2.6 Gy) and test dose (1.2 Gy) at five different preheat temperatures (three aliquots for each preheat temperature). After four cycles all aliquots were stored in the oven at 100 1C, following exposing to 2.6 Gy dose. After 3 weeks storage, the regeneration signal and response to the test dose (1.2 Gy) was measured (fifth cycle). Then two more cycles of the SAR procedure with a fixed regeneration (2.6 Gy) and test dose (1.2 Gy) was repeated. The results are shown in Fig. 4. The average of these measurements at different preheat temperature shows a drop at cycle number 5. The fading ratio was calculated by the ratio of sensitivity corrected IRSL of the stored dose (L5/T5) divided by the average of sensitivity corrected IRSL before storage (L4/T4) and the ratio of prompt measurement after storage (L6/T6). This suggests that the sample suffers from 10% fading (Fig. 4). 6 7 8 4 5 Cycle no. 1.2 1 0.8 0.6 0.4 0.2 0 0 Average 1 2 3 4 5 Cycle no. Fig. 4. Fading test: sensitivity corrected IRSL signal as a function of repeated cycles using different preheating temperature shown in the figures. The bottom right is the average of 15 aliquots for each cycle. ARTICLE IN PRESS M. Fattahi, R.T. Walker / Quaternary Geochronology 2 (2007) 284–289 After fading correction. Uranium, thorium and potassium concentrations were measured using field gamma spectrometry. Present-day moisture contents were determined by drying at 40 1C in the laboratory. The conversion factors for water contents of Aitken (1985) were used for the calculation of alpha, beta and gamma dose rates. Alpha and beta dose rates were corrected for attenuation due to grain size using the factors of Bell (1980) and Mejdahl (1979). b a 1.7770.29 2.870.3 1.5870.20 1.1470.04 0.4470.04 0.014770.000 2.77970.160 0.87170.050 1.5 90–150 S6 2.5 0.85270.051 U (ppm) K (%) Depth (m) Water (%) Grain (mm) Sample Table 2 Values used to calculate luminescence ages from Sabzevar fault, NE Iran Th (ppm) Cosmic (Gy/ka) Dint (mGy/yr) DExt (mGy/yr) Dose rate (mGy/yr) Dea (Gy) Ageb (ka) 288 4. Discussion and conclusions The linear relationship between Lx and Tx which passes through the origin in third method (Table 1, Fig. 1c) satisfies the basic requirement of SAR method. For two other methods an increasing or decreasing intercept (in comparison to zero) can be interpreted that Lx and Tx show different sensitivity changes (Fig. 1a, b). All methods recovered the known laboratory dose in all preheat temperature examined within their estimated error and the best result was obtained by the second method (Fig. 2b). However, surprisingly only the second method has shown a preheat plateau for the natural De (Fig. 3b). The first method has shown a rising trend of De with increasing preheat temperature (Fig. 3a). Some one may suggest that the rising trend can be the result of thermal transfer. Preheating can result in thermal transfer from shallow traps to the traps sampled during OSL measurement. Unwanted thermal transfer can occur in nature if the light-insensitive traps are thermally unstable, and part of their charge is re-trapped in the OSL trap. Such thermal transfer can cause an overestimation of age and cannot be avoided by using a low preheat. However, there is no evidence that this sample is suffering from thermal transfer. The third method has shown two preheat plateau. One plateau is shown at around 1.85 Gy between 210 and 250 1C and the other at around 2.54 Gy between 270 and 290 1C. However, based on dose recovery test and a clear plateau in the second method for De values (mean De ¼ 2.5470.2 Gy) in the preheat temperature range of 230–290 1C, we used this mean value for age determination. The result of fading tests showed that the feldspar grains are subject to anomalous fading (10%) and as such, have provided an apparent deposition age which is younger than the real age. If we consider a natural fading of 10% then the De can increase to 2.80 Gy and the age can increase to 1.770.3 ka. Table 2 shows the values used to determine sample age and the derived age estimate. External dose was measured by a portable gamma spectrometer and for calculation of internal K dose rate, potassium contents of 12.570.5% were used (see Preusser, 2003). The IRSL age of feldspar grains therefore indicates that the colluvial sediments are young, and that the faulting that cuts them must date from less than 1800 years ago. Given this age range, it is likely that the faulting observed in Fig. S2 does represent surface deformation from the AD 1052. Baihaq earthquake, which is therefore likely to have occurred on the Sabzevar fault. Acknowledgements This study has been partly supported by the Research Department of University of Tehran in the form of a project (6201002/1/01) to MF and partly by the Royal Society of London in the form of an award (2004/R3-RW) to MF and RTW. The Oxford University Centre of ARTICLE IN PRESS M. Fattahi, R.T. Walker / Quaternary Geochronology 2 (2007) 284–289 Environment has provided all the luminescence experimental facilities and requirements. Logistical help was provided by the Geological Survey of Iran and we thank them for their continued support of our work in Iran. We especially thank Morteza Talebian, Abbas Bahroudi, Hamid Nazari and Manuchehr Ghorashi. RTW is supported by NERC and the NERC-funded Centre for the Observation and Modelling of Earthquakes and Tectonics (COMET). Editorial handling by: R. Roberts Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version at doi:10.1016/j.quageo. 2006.06.006. References Aitken, M.J., 1985. Thermoluminescence Dating. Academic Press, London. Ambraseys, N.N., Melville, C.P., 1982. A History of Persian Earthquakes. Cambridge University Press, UK. Bell, W.T., 1980. Alpha dose attenuation in quartz grains for thermoluminescence dating. Ancient TL 12, 4–8. Blair, M.W., Yukihara, E.G., McKeever, S.W.S., 2005. Experience with single aliquot OSL procedures using coarse-grain feldspars. Radiation Measurement 39, 361–374. Fattahi, M., 2001. Studies on red thermoluminescence and infrared stimulated red luminescence. Unpublished D.Phil. Thesis, Oxford University, Oxford. 289 Fattahi, M., Stokes, S., 2004. Absorbed dose evaluation in feldspar using a single-aliquot regenerative-dose (SAR) infrared-stimulated red luminescence protocol. Radiation Measurements 38 (1), 127–134. Fattahi, M., Walker, R., Hollingsworth, J., Bahroudi, A., Nazari, H., Talebian, M., Armitage, S., Stokes, S., 2006. Holocene slip-rate on the Sabzevar thrust fault, NE Iran, determined using optically-stimulated luminescence (OSL). Earth and Planetary Science Letters 245 (3–4), 673–684. Hutt, G., Jaek, I., Tchonka, J., 1988. Optical dating: Kfeldspars optical response stimulation spectrum. Quaternaly Science Reviews 7, 381–386. Mejdahl, V., 1979. Thermoluminescence dating: beta-dose attenuation in quartz grains. Archaeometry 21, 61–72. Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements 32 (1), 57–73. Porat, N., Wintle, A.G., Amit, R., Enzel, Y., 1996. Late Quaternary earthquake chronology from luminescence dating of colluvial and alluvial deposits of the Arava Valley, Israel. Quaternary Research 46, 107–117. Preusser, F., 2003. IRSL dating of K-rich feldspars using the SAR protocol: comparison with independent age control. Ancient TL 21 (1), 17–23. Rhodes, E.J., 2000. Observations of thermal transfer OSL signals in glacigenic quartz. Radiation Measurements 32, 595–602. Talebian, M., Fielding, E.J., Funning, G.J., Jackson, J., Nazari, H., Parsons, B., Priestley, K., Qorashi, M., Rosen, P.A., Walker, R., Wright, T.J., 2004. The 2003 Bam (Iran) earthquake—rupture of a blind strike-slip fault. Geophysical Research Letters 31, L11611. Wallinga, J., Murray, A., Duller, G., 2000a. Underestimation of equivalent dose in single-aliquot optical dating of feldspars caused by preheating. Radiation Measurements 32, 691–695. Wallinga, J., Murray, A., Wintle, A., 2000b. The single aliquot regenerative-dose (SAR) protocol applied to coarse-grain feldspar. Radiation Measurements 32, 529–533.