Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

New evidence for large earthquakes on the Central Iran plateau: palaeoseismology of the Anar fault

Geophysical Journal International, 2012
...Read more
For Peer Review                                                                                                                                Geophysical Journal International hal-00681707, version 1 - 14 May 2012 Author manuscript, published in "Geophysical Journal International 189 (2012) 6-18" DOI : 10.1111/j.1365-246X.2012.05365.x
For Peer Review New evidence for large earthquakes on the Central Iran plateau: Paleoseismology 1 of the Anar fault 2 Foroutan 1,2,3 M., M. Sébrier 2,3 , H. Nazari 4 , B. Meyer 2,3 , M. Fattahi 5,6,7 , A. Rashidi 8 , K. Le 3 Dortz 2,3,* and M. D. Bateman 6 4 1: Geological Survey of Iran, Azadi square, Meraj avenue, PO Box: 13185-1494, Tehran, Iran 5 email: foroutan@gsi.ir 6 2: UPMC Univ Paris 06, ISTEP UMR 7193, Université Pierre et Marie Curie, F-75005, Paris, France 7 email: mohammad.foroutan@upmc.fr, bertrand.meyer@upmc.fr, michel.sebrier@upmc.fr, 8 3: CNRS, ISTEP, UMR 7193, F-75005, Paris, France 9 4: Research Institute for Earth Sciences, Geological Survey of Iran, Po Box: 13185-1494,Tehran, Iran 10 email: h.nazari@gsi.ir 11 5: The Institute of Geophysics, University of Tehran, Tehran, Iran 12 email: mfattahi@ut.ac.ir 13 6: Sheffield Centre for International Dryland Research, Department of Geography, University of 14 Sheffield, Winter Street, Sheffield S10 2TN, UK 15 email: M.D.Bateman@Sheffield.ac.uk 16 7: The School of Geography, University of Oxford, OX1 3QY, UK 17 email: morteza.fattahi@ouce.ox.ac.uk 18 8: Geological Survey of Iran, Postal Code 7615736841, Kerman, Iran 19 email: kermangeo@gsi-iran.org 20 21 * Now at: Laboratoire de Géologie, ENS, UMR 8538, 75231 Paris, ledortz@geologie.ens.fr 22 23 Abstract 24 The Central Iran plateau appears aseismic during the last few millenniums based on 25 instrumental and historical seismic records. Nevertheless, it is sliced by several strike-slip 26 faults that are hundreds kilometres-long. These faults display along-strike, horizontal offsets 27 of intermittent gullies that suggest the occurrence of earthquakes in the Holocene. 28 Establishing this is crucial for accurately assessing the regional seismic hazard. The first 29 paleoseismic study performed on the 200-km-long, NS striking Anar fault shows that this 30 right-lateral fault hosted three large (M w ァ earthquakes during the Holocene or possibly 31 Uppermost Pleistocene for the older one. These three seismic events are recorded within a 32 sedimentary succession, which is not older than 15 ka, suggesting an average recurrence of at 33 Page 1 of 28 Geophysical Journal International 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 hal-00681707, version 1 - 14 May 2012
Author manuscript, in "Geophysical Journal International 189 (2012) 6-18" Geophysical Journal published International DOI : 10.1111/j.1365-246X.2012.05365.x     Fo     rP               ev rR ee   w ie hal-00681707, version 1 - 14 May 2012   Page 1 of 28 New evidence for large earthquakes on the Central Iran plateau: Paleoseismology 2 of the Anar fault 3 Foroutan1,2,3 M., M. Sébrier2,3, H. Nazari4, B. Meyer2,3, M. Fattahi5,6,7, A. Rashidi8, K. Le 4 Dortz2,3,* and M. D. Bateman6 5 6 1: Geological Survey of Iran, Azadi square, Meraj avenue, PO Box: 13185-1494, Tehran, Iran email: foroutan@gsi.ir 7 8 2: UPMC Univ Paris 06, ISTEP UMR 7193, Université Pierre et Marie Curie, F-75005, Paris, France email: mohammad.foroutan@upmc.fr, bertrand.meyer@upmc.fr, michel.sebrier@upmc.fr, 9 3: CNRS, ISTEP, UMR 7193, F-75005, Paris, France 10 11 4: Research Institute for Earth Sciences, Geological Survey of Iran, Po Box: 13185-1494,Tehran, Iran 12 13 5: The Institute of Geophysics, University of Tehran, Tehran, Iran 14 15 16 6: Sheffield Centre for International Dryland Research, Department of Geography, University of 17 18 7: The School of Geography, University of Oxford, OX1 3QY, UK 19 20 21 22 8: Geological Survey of Iran, Postal Code 7615736841, Kerman, Iran email: h.nazari@gsi.ir email: mfattahi@ut.ac.ir email: morteza.fattahi@ouce.ox.ac.uk ev email: kermangeo@gsi-iran.org rR ee Sheffield, Winter Street, Sheffield S10 2TN, UK email: M.D.Bateman@Sheffield.ac.uk * Now at: Laboratoire de Géologie, ENS, UMR 8538, 75231 Paris, ledortz@geologie.ens.fr Abstract iew 23 24 rP hal-00681707, version 1 - 14 May 2012 1 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Geophysical Journal International 25 The Central Iran plateau appears aseismic during the last few millenniums based on 26 instrumental and historical seismic records. Nevertheless, it is sliced by several strike-slip 27 faults that are hundreds kilometres-long. These faults display along-strike, horizontal offsets 28 of intermittent gullies that suggest the occurrence of earthquakes in the Holocene. 29 Establishing this is crucial for accurately assessing the regional seismic hazard. The first 30 paleoseismic study performed on the 200-km-long, NS striking Anar fault shows that this 31 right-lateral fault hosted three large (Mw§  earthquakes during the Holocene or possibly 32 Uppermost Pleistocene for the older one. These three seismic events are recorded within a 33 sedimentary succession, which is not older than 15 ka, suggesting an average recurrence of at Geophysical Journal International most 5 ka. The six OSL ages available provide additional constraints and allow estimating 35 that the three earthquakes have occurred within the following time intervals: 4.4±0.8 ka, 36 6.8±1 ka, and 9.8±2 ka. The preferred age of the more recent event, ranging between 3600 yrs 37 and 5200 yrs, suggests that the fault is approaching the end of its seismic cycle and the city of 38 Anar could be under the threat of a destructive earthquake in the near future. Additionally, our 39 results confirm a previous minimum slip rate estimate of 0.8±0.1 mm/yr for the Anar 40 fault indicating the westernmost prominent right‐lateral faults of the Central Iran 41 plateau are characterized by slip rates close to 1 mm/yr. These faults, which have 42 repeatedly produced destructive earthquakes with large magnitudes and long 43 recurrence interval of several thousands of years during the Holocene, show that the 44 Central Iran plateau does not behave totally as a rigid block and that its moderate 45 internal deformation is nonetheless responsible for a significant seismic hazard. 46 Introduction ee 47 rP 48 The Central Iran Plateau is a wide region experiencing low GPS deformation rates 49 and is commonly described as a rigid block (e.g., Jackson & McKenzie, 1984; Vernant et 50 al., 2004). The region (Figure 1) is nonetheless sliced by several strike‐slip faults with 51 clear morphological traces (Walker & Jackson, 2004; Meyer et al., 2006; Meyer & Le 52 Dortz, 2007) that contrast with the very few earthquakes recorded in the region 53 (Ambraseys & Jackson, 1998). Destructive earthquakes have occurred close to or along 54 the Lut faulted borders only, and, according to the historical and instrumental records 55 (Ambraseys & Melville, 1982; Ambraseys & Jackson, 1998), the prominent right‐lateral 56 strike‐slip faults inland remained quiescent for millenniums. Although the absence of 57 historical record of earthquakes in remote and uninhabited desert does not mean an 58 absence of earthquakes, knowledge of the behaviour of such faults and assessment of 59 the regional seismic hazard requires paleoseismic studies and depends on the selection 60 of suitable trenching sites. This is the case for the region of the Dehshir and Anar faults. 61 While a recent paleoseismic study stated the occurrence of large and infrequent 62 earthquakes on the Dehshir fault (Nazari et al., 2009; Fattahi et al., 2010), the seismic 63 behaviour of the Anar fault is still to be assessed. iew ev rR hal-00681707, version 1 - 14 May 2012 34 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 2 of 28 Page 3 of 28 The Anar fault is a 200‐km‐long NS strike‐slip fault located nearby 55°E, in the 65 middle of the Central Iran Plateau (Figure 1). The northern part of the fault is located 66 within a mountainous region where several closely spaced splays cut across the Kuh-e 67 Kharanaq range. The southern portion of the fault runs along the Kuh-e Bafq range and cuts 68 right across the western piedmont of the range and across the Anar salt flat. The fault goes 69 through the populated city of Anar and further bends eastwards to reactivate a thrust fault 70 to the south. According to 71 luminescence (OSL) dating of cumulative offset of alluvial fans, the southern portion of the 72 fault slips at a minimum rate of 0.8 mm/year (Le Dortz et al., 2009). Both the sharpness 73 of these cumulative offsets and the absence of along fault microseismicity suggest these 74 offsets have accrued through large and infrequent earthquakes rather than by creeping. 75 At a few places, the offset of small gullies that ranges between 2.5 and 3.5 m (Figure 2) 76 may be interpreted as the amount of coseismic slip during the last earthquake but clear 77 evidence for a continuous and distinctive surface break is missing. Trenching appears 78 therefore necessary to document earthquakes and access the Recent seismic history of 79 the fault. rR 81 Be cosmic ray exposure (CRE) and optically-stimulated ee 80 10 rP Site selection and trench stratigraphy ev 82 We scrutinized the fault on Quickbird satellite imagery and in the field to select 83 the most favourable place to conduct paleoseismic investigations. The selected site is 84 located 35 km north of the city of Anar along a very clear portion of the N175°E fault 85 trace that is highlighted by an E‐facing scarp (Figure 3). This fault scarp is readily seen 86 cutting across numerous rills and ephemeral streams that incise a fan surface 87 characterized by a subdued bar‐and‐swale morphology. The abandoned alluvial fan 88 surface comprises a loose desert pavement of varnished clasts separated by sandy‐silty 89 material. The smooth scarp has a minimum height of 35 cm near the excavation (Figure 90 3b). South of the trench site, the scarp height increases progressively to reach one meter 91 where small streams, which have been dammed and channelled along the fault, 92 rejuvenate this scarp. North of the trench site, the scarp height also increases, but the 93 streams are wider and deeper than to the south and have been able to maintain their 94 courses to flow across the fault trace. At the trench site, the possibility of damming small 95 intermittent streams during the emplacement of the fan material is high and the iew hal-00681707, version 1 - 14 May 2012 64 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Geophysical Journal International Geophysical Journal International possibility of subsequent erosion low, maximizing the chances to record the earthquakes 97 coeval with alluvial aggradation and subsequent colluvial deposition. The next 98 paragraphs report the observations gathered at the selected site, where trenching 99 allowed us to distinguish three unambiguous events. 100 The eastern tip of the excavated trench is located about 5 km west of the Kuh‐e 101 Bafq range at 31.1953°N; 55.1536°E, and 1544‐m‐altitude above sea level (Figures 1 and 102 3b). The trench strikes N85°E, it has a 28‐m‐length, a depth between 4.3 and 4.7 m, and 103 a width of some 1.5 m (Figures 3, 4, and 5). Trench walls expose a total deposit thickness 104 of 5.9 m with fairly flat, 0.1‐to 1‐m‐thick beds that correspond to medium‐distal, alluvial 105 fan facies. Trench stratigraphy is relatively straightforward as these beds are continuous 106 and easily correlated across the fault zone located approximately in the middle of the 107 trench (Figure 5). Overall the beds found in the trench are composed of sands and 108 gravels mixed with variable amounts of silt and clay, these deposits correspond mostly 109 to a debris‐flow dominated alluvial fan. Although debris‐flows and sheet‐floods appears 110 to dominate, sediment dynamics is not easy to determine precisely for each bed due to 111 the medium‐distal deposit conditions. rR ee rP 112 A total of 17 units have been defined in the excavated trench (see Table 1 and 113 Figures 4 and 5). The uppermost part of the trench, approximately the last meter, is 114 made of the thinner units 13 to 17 containing smaller pebbles than the alluvial units 1 to 115 8 and an almost clay‐free matrix contrary to the units 9 to 11. This last meter hence is 116 made of runoff deposits. The only evidence for a significant break in alluvial 117 sedimentation is the γ0.15‐m‐thick, gypsiferous calcrete that is located toward the top of iew ev hal-00681707, version 1 - 14 May 2012 96 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 118 unit 11. This latter unit and units 9 and 10 correspond to reddish brown, mud‐ 119 dominated debris flows. OSL was used to date the alluvial layers seen in the trench. 120 121 OSL dating 122 Six lenses of fine sandy‐silts intercalated between fanglomerates at various depths 123 below the surface were sampled for OSL in opaque tubes (Figure 5). The Single Aliquot 124 Regeneration (SAR) protocol (e.g.; Murray and Wintle, 2000) was employed for the 125 Equivalent dose (De) measurement once quartz had been extracted and cleaned from each 126 sample. The analytical procedures employed are identical to that applied to similar samples Page 4 of 28 Page 5 of 28 from the neighbouring Sabzevar, Doruneh, and Dehshir faults (Fattahi et al, 2006; 2007; 128 2010; Le Dortz et al., 2011). In order to make all data consistent, three samples with ages that 129 were previously published in Le Dortz et al. (2009), including one sample collected 130 approximately 8 km farther north (OSL-2 on Figure 1 and in Table 2), have been refined in 131 the light of procedural development outlined by Fattahi et al. (2010). Table 2 provides the 132 relevant information for OSL ages in years from present (2010) with 1 sigma errors. 133 Initial attempts to use single grains for De determination failed due to the dimness of 134 OSL signal. As a result, De measurements were undertaken on 9.6-mm-diameter aliquots 135 containing approximately 1500-2000 grains. Whilst normally this might result in averaging 136 out of any multiple dose component within a sample, here it is assumed that for these dim 137 samples the luminescence signal from each aliquot was produced by a relatively few number 138 of bright grains and thus may be considered as almost measuring at single grain level. 139 Therefore, the De distribution of the single aliquot De measurements is considered to be 140 almost a true reflection of the actual De distribution within a sample. For samples showing 141 unimodal, apparently normally distributed De’s with a low over-dispersion (Ant-I and Ant-IV 142 in Table 2), which are interpreted as having had the OSL signal reset (bleached) prior to 143 burial, Central Age Model (CAM) was employed for calculation purposes. For the remaining 144 samples (Ant-II, Ant-III, Ant-V, Ant-VI and OSL-2 in Table 2), the depositional setting, field 145 sedimentary logs, and the scatter of the replicate aliquot De data indicated that prior to burial, 146 full resetting (bleaching) of the OSL signal had not taken place and/or that the sediments had 147 undergone some post-depositional disturbance (Bateman et al., 2007). Indeed, the small size 148 of the catchment area at the trench site, less than 20 km2, is indicative of a rapid transport 149 before the emplacement of the fan (Le Dortz et al., 2009). The great majority of the Anar 150 trench deposits come from high-discharge depositional events with limited surface exposure, 151 which favour partial bleaching (Rittenour, 2008). As a consequence, Finite Mixture Model 152 (FMM; Roberts et al. 2000) was used where samples showed skewed, scattered, or 153 multimodal distributions and the dominant De component was used for age calculations (as in 154 Fattahi et al., 2010). This approach yielded ages in accordance with site stratigraphy except 155 for sample Ant-III whose age was too young. This sample has exhibited the highest over- 156 dispersion value (51%, Table 2) of all measured samples and three De component were 157 extracted by FMM for this sample, the highest representing 30% of the signal and the smallest 158 47%. It is possible to get an age that fits with stratigraphy using the higher two De 159 components extracted using FMM, corresponding to ages of 7.3±0.5 ka and 10.7±0.8 ka, iew ev rR ee rP hal-00681707, version 1 - 14 May 2012 127 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Geophysical Journal International Geophysical Journal International respectively. However, there are no good reasons to accept components representing only 161 23% or 30% of the data over the 47% of the data incorporated into the dominant smallest 162 FMM component. The latter cannot be ignored on the basis of partial bleaching as it has a 163 lower De and is unlikely to represent post-depositional disturbance as it is the dominant De 164 component but also based on observed bedding within the unit. However, unlike most other 165 samples studied, Ant-III was sampled actually on a stratigraphic boundary (between units 8 166 and 9). As the dose-rate for this sample is based only on the activity from unit 9, a difference 167 in the gamma radiation dose received from unit 8 might help explain the apparent under- 168 estimation of age based on the lowest FFM De component. Unfortunately field-based gamma- 169 spectrometer readings or material for analyzing the radioactive elements within unit 8 were 170 not available to test this and so the age of Ant-III has not been included within the subsequent 171 site interpretation. Thus, the ages of the trench units have been constrained based on only five 172 OSL ages (Ant-I, Ant-II, Ant-IV, Ant-V, and Ant-VI). 173 174 Ages of trench units ee rP 175 The samples collected within the trench define a time range spanning from 5.8 to 176 14.9 ka. The oldest age Ant‐I (13.6±1.3 ka) stands at the bottom of the trench and 177 belongs to one of the oldest stratigraphic units (unit 2), it indicates that all the trench 178 units should be younger than 15 ka. As the youngest OSL sample Ant‐VI (6.2 ± 0.4 ka) is 179 located at 0.8 m below the surface (unit 14), these ages define a maximum time interval 180 between 0 and 14.9 ka and a minimum one between 6.6 and 12.3 ka, indicating the 181 trench deposits aggraded during the uppermost Marine Isotopic Stage 2 (MIS‐2γͳʹ‐ 182 22 ka) and part, or the entirety of MIS‐1. The age of the fan surface at the trench site 183 appears poorly constrained between 0 and 6.6 ka. iew ev rR hal-00681707, version 1 - 14 May 2012 160 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 184 A proxy to the surface age may be obtained by estimating average sedimentation 185 rates. Assuming this fan surface is still active, the whole 5.9 m of sedimentary units seen 186 in the trench should have been aggraded during the last 14.9 or 12.3 ka at an overall 187 sedimentation rate ranging between 0.4 and 0.48 mm/yr. Deposit thicknesses above the 188 samples Ant‐I (13.6±1.3 ka) and Ant‐II (12.4±0.6 ka) yield slightly lower rates ranging 189 between 0.28 and 0.36 mm/yr. Conversely, if one assumes the top of the fan surface was 190 abandoned just after 6.6 ka, the oldest possibility for the youngest sample (Ant‐VI), the Page 6 of 28 Page 7 of 28 sedimentation rate may have reached 1 mm/yr. These two extreme hypotheses raise the 192 problem of the abandonment age of the alluvial fan at the precise location where the 193 trench was excavated. The refined age of OSL‐2 (10.1±0.6 ka versus 11.8 ±6.5 ka in Le 194 Dortz et al., 2009), which is sampled some 8 km north of the trench (see location on 195 Figure 1, see also Figures 4a and 7 in Le Dortz et al., 2009) at 0.8 m below the surface of 196 197 ƒ‘–Š‡” ˆƒ ‘ –Š‡ ”‹•‡” ‘ˆ ƒ γͶǤͷ  ‹ ‹•‡† †”› •–”‡ƒ, suggests its surface was 198 be older than 6.6 ka. These ages are in agreement with the late Pleistocene and Holocene 199 regional climate scenario, for the central and eastern Iran, as recently proposed by 200 Walker & Fattahi (2011). In fact, the satellite image in the vicinity of the trench site 201 (Figure 3a) shows this location corresponds to an area where the fan surface has been 202 exceptionally preserved from the backward erosion of deeply incised dry streams. 203 Therefore, the late aggradation on the fan surface at the trench site (‫׽‬1544 m a.s.l.), 204 resulting from surface runoff, lasted until more recently than some 8 km further north 205 on a more proximal alluvial fan (‫׽‬1748 m a.s.l.) where Le Dortz et al. (2009) collected 206 the sample OSL‐2. In conclusion, the best estimate for the aggradation rate of the trench 207 sediments is provided by the depth difference of 3.3 m between Ant‐I (13.6 ± 1.3 ka) and 208 Ant‐VI (6.2 ± 0.4 ka) samples, this rate ranges between 0.36 and 0.58 mm/yr. 209 Interestingly, this rate is comparable to the late Pleistocene‐Holocene aggradation rates 210 derived from alluvial sediments in similar deposition settings at southwestern Nebraska, 211 Negev desert and Central Iran plateau (Daniels et al., 2003; Guralnik et al., 2011; 212 Schmidt et al., 2011). Consequently, the best averaged net sedimentation rate, if 213 meaningful in such environment, amounts to 0.47±0.11 mm/yr suggesting that the final 214 aggradation of the trench units ended between 3.6 and 5.2 ka to the west of the fault 215 zone and approximately one thousand years later to the east of the fault zone. The 216 abandonment age of the fan surface at the trench site (4.4±0.8 ka) is comparable with 217 late Holocene drought cycles at 4.2 ka, proposed by Staubwasser et al. (2003) to the 218 southeast of the Persian Gulf. abandoned some 9‐10 ka ago at this site while the fan surface at the trench site cannot iew ev rR ee rP hal-00681707, version 1 - 14 May 2012 191 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Geophysical Journal International 219 220 Seismic event identification 221 The surface geomorphology, particularly, the characteristics of the fault scarp, 222 detailed geological and structural analyses of the trench walls (bed unconformities, Geophysical Journal International sealed fault strands, and fissure fills or sand‐blows), and restoration of the trench log 224 permit to trace three individual event horizons that correlate with three destructive past 225 earthquakes, which ruptured the ground surface. The evidence for these three seismic 226 events, which are labelled A, B, and C from the youngest to the oldest, is presented 227 below. 228 Event A, the youngest event is responsible for the 35‐cm‐high, E‐facing scarp, 229 which is observed at the surface in the vicinity of the trench (Figures 3b and 3c). The 230 trench log shows this scarp is located above a nearly 1‐m‐wide, steep faulted zone 231 associated with an E‐facing flexure of the upper trench units; i.e., downthrown to the 232 east (Figure 5). Units 16 and 17 are not observed west of the fault zone nor warped 233 eastward. They rest unconformably on unit 15 dipping about 25°E (Figure 6), hence 234 indicating the faulting and flexuring deformation postdate unit 15 and predate units 16 235 and 17. As there are no significant thickness differences of the upper trench units on 236 both sides of the fault zone, the faulting and flexuring deformation appear to result from 237 a sudden and discrete slipping event on the fault zone instead of a continuous shear; i.e., 238 creep slip (Figure 7‐step 1). Inasmuch as the upper trench units are seen faulted up to 239 unit 15, the discrete slipping event should relate to an earthquake that ruptured the 240 surface producing the scarp. The event horizon of this earthquake locates on the top 241 boundary of unit 15, so that it stands between this latter and units 16 and 17 to the east 242 of the fault and corresponds approximately with the present‐day topographic surface to 243 the west of the fault. This suggests that the western block has been slightly eroded as 244 well as partly concealed by units 16 and 17 so that its present‐day height 245 underestimates its original height at the time of the earthquake (Figure 7‐step 2). Then, 246 unit 17 is a local sag pond deposit caused by event A. iew ev rR ee rP hal-00681707, version 1 - 14 May 2012 223 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 247 Backstripping of event A suggests its vertical throw is of some 60 cm (Figure 7‐ 248 step 3), confirming that the 35‐cm high scarp at the surface has been partly degraded. 249 Taking into account the Anar fault is dominantly a strike‐slip fault (Figure 2), the 60 cm 250 of vertical displacement should be associated with at least a couple of meters of right‐ 251 lateral displacement. Then, the magnitude of event A should be at least on the order of 252 Mw 7. The age of event A is poorly constrained as the uppermost OSL sample (Ant‐VI), 253 which is at 0.8‐m‐depth, comes from unit 14. This indicates only that event A is younger 254 than 6.6 ka. Tentatively, the sediment thickness, which stands between the sample Ant‐ Page 8 of 28 Page 9 of 28 VI and the event horizon of event A, may be used to propose a rough estimate of the 256 youngest possible age for this seismic event based on the aggradation rate between Ant‐ 257 VI and Ant‐I. Since aggradation of unit 15 ended between 3600 and 5200 years (see 258 previous section) event A should not be younger than 3600 years and might be as old as 259 5200 years. This time interval will be considered as the preferred age of event A in the 260 following discussion. 261 Event B, the penultimate event, is easily determined from analysis of the trench 262 walls. Several filled fissures (unit 12), which postdate unit 11 and predate unit 13, are 263 seen (Figures 5 and 7‐step 4). A careful analysis of these filled fissures permits to 264 conclude they correspond indeed to either open cracks or sand‐blows (Figures 8a and 265 8b). It is possible to observe these features formed suddenly as they disrupt the hard 266 gypsiferous calcrete (e.g., Figure 8b), which is located in the top part of unit 11. Thus, the 267 interface between units 11 and 13 corresponds to the event horizon of event B. The 268 backstripping analysis of the trench log (Figure 7), removing the effects of event A and 269 the units aggraded between events A and B, allow reconstructing the event B horizon 270 just after the penultimate earthquake (Figure 7‐step 4). This reconstruction suggests 271 that the rupture of event B should have occurred on two parallel fault strands of the 272 Anar fault zone, totalling a vertical throw of some 40 cm, distributed into 25 cm on the 273 eastern strand and 15 cm on the western one. Thus, the penultimate earthquake appears 274 to have a vertical throw close to that of the last one, hence a magnitude lightly 275 comparable with the one of event A. The age of event B is well constrained by three OSL 276 ages (Ant‐V, ‐IV, and ‐VI) and bracketed between 5.8 ka and 7.8 ka yielding an average 277 age of 6.8±1 ka for this earthquake (Figure 9). iew ev rR ee rP hal-00681707, version 1 - 14 May 2012 255 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Geophysical Journal International 278 Event C, the oldest seismic event, can be safely identified on the trench walls. 279 Evidence for a third earthquake rests on the facts that lower trench units (5 to 8) exhibit 280 higher vertical throws than middle ones (9 to 11). Different reconstructions have been 281 tested to restore the trench lower units, the best fit is obtained when considering the 282 interface between units 8 and 9 as the event horizon of a surface rupturing earthquake 283 (Figure 7‐step 6). This reconstruction favours that event C ruptured along the eastern 284 fault strand with a vertical displacement of about 25 cm. Nevertheless, this vertical 285 displacement is poorly constrained and higher estimates might be obtained according to 286 the amount of erosion chosen (dashed line on Figure 7‐step 6). Even if this vertical Geophysical Journal International throw is poorly determined and appears smaller than the ones of the subsequent 288 earthquakes, this does not imply a much smaller magnitude as the amount of coseismic 289 vertical displacement is highly variable along a strike‐slip fault (e.g., Barka, 1996; Barka 290 et al., 2002). The age of event C is not well constrained, it occurred prior to the OSL age 291 of sample Ant‐V (7.1±0.7 ka) and after that of Ant‐II (12.4±0.6 ka) so that it is bracketed 292 within the interval between 6.4 and 13 ka, with a minimum interval of 4 ka (Figure 9). 293 We have investigated the possibility for a fourth event in the lowermost part of the 294 trench (Figure 5). However, the lowermost units (1 to 4) cannot be followed on the 295 eastern side of the trench precluding the possibility to document an additional event. 296 297 Discussion and Conclusions rP 298 Due to the reduced size of the surface sag pond, a 3D trench exploration was not 299 undertaken at the trench site so that the slip per event was not directly measurable for 300 the three identified seismic events, A, B, and C. Accounting for the lack of morphological 301 segmentation along the Anar fault (Le Dortz et al., 2009) and the excavation of a single 302 trench, there are no possibilities to calculate confidently any magnitude. The restored 303 vertical throws per event suggest only that the related magnitudes are more likely on 304 305 the order of Mwγ͹ (see previous section). The right‐lateral offsets observed at the 306 streams are seen offset by 8±0.5 m (Le Dortz et al., 2009 and Figure 3a) while several 307 weakly incised small dry gullies show offsets of 3±0.5 m only (Figure 2). Considering 308 these lower offsets were more likely formed during the most recent earthquake, they 309 provide the best estimate for the coseismic horizontal slip of event A. Such a coseismic 310 slip at the surface agrees with a magnitude close to 7. Thus the higher offsets, which are 311 only observed where the alluvial surface is older, should represent the cumulative 312 horizontal slip of the last three events that are seen in the trench. Therefore, the three 313 events, which are observed within the trench to the north of Anar city, should have had 314 slip per event on the order of 3 m, suggesting they are of similar magnitudes. ev rR ee surface (Figures 2 and 3a) permit to reinforce this inference. Deeply incised intermittent iew hal-00681707, version 1 - 14 May 2012 287 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 10 of 28 315 Overall, the new data presented in this paper give evidence for at least three 316 seismic events of similar sizes on the Anar fault (Figure 9). OSL age of Ant‐I sample 317 indicates that the aggradation of the trench units did not started much before 14.9 ka or Page 11 of 28 to the latest by 12.3 ka. Since only 3 events have occurred during the last 15 ka, this 319 provides a rough estimate of at most 5 ka for the average maximum time interval 320 between two earthquakes. The youngest age possibility may reduce this average time 321 322 interval between two earthquakes –‘ γͶ ka. Considering the best estimates of average 323 respectively (Figure 9), the time interval between two subsequent earthquakes is ill‐ 324 defined and might vary significantly. Nevertheless, the average ages for events A, B, and 325 C indicate that the intervals between two subsequent earthquakes should be of 2400 326 and 2900 years between events A‐B and B‐C, respectively. As the elapsed time since the 327 last earthquake is 3600 years at least and 5200 years at most, this suggests we are 328 getting close to the end of the seismic cycle and may anticipate a destructive earthquake 329 for the Anar city in the near future. ages for the three seismic events A, B, and C are: 4.4±0.8 ka, 6.8±1 ka, and 9.7±3.3 ka, rP 330 Finally, the cumulative offset of 8±0.5 m (Le Dortz et al., 2009) postdating fan 331 aggradation and the refined age of fan abandonment given by OSL‐2 sample 332 (10.1±0.6 ka) confirms a minimum slip rate estimate of 0.8±0.1 mm/yr for the Anar 333 fault. Then, the westernmost prominent right‐lateral faults of the Central Iran plateau, 334 namely the Dehshir and Anar faults, which are active though void of historical and 335 instrumental earthquakes, are characterized by slip rates close to 1 mm/yr (Le Dortz et 336 al., 2009 and 2011). Such faults have repeatedly produced destructive earthquakes with 337 large magnitudes (Mwγ͹) and long recurrence interval of several thousands of years ev rR ee 338 during the Holocene (Nazari et al., 2009; Fattahi et al., 2010 and this paper). This 339 demonstrates that the Central Iran plateau does not behave totally as a rigid block and 340 that its moderate internal deformation is nonetheless responsible for a significant 341 seismic hazard. iew hal-00681707, version 1 - 14 May 2012 318 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Geophysical Journal International 342 343 Acknowledgements 344 This study benefited of the logistic and financial assistance from Geological 345 Survey of Iran. Université Pierre et Marie Curie and INSU‐CNRS provided 346 complementary funding for the fieldwork and OSL measurements. M. Foroutan 347 acknowledges a grant from the French Embassy in Tehran for part of his PhD thesis and 348 complementary support from UPMC‐ISTeP. David Lambert, Attaché Scientifique et Geophysical Journal International Culturel, is thanked for his continuing support to the cooperation between UPMC and 350 GSI. M. Hosseini from Kerman GSI office helped with the logistics of the fieldwork. B. 351 Oveisi, M. Nazem Zadeh, A. Agha Hosseini, and M.A. Shokri are thanked for various 352 contributions to the fieldwork, H. Bani Assadi and M. Adhami for their safe driving in the 353 field. M. Fattahi would like to thank the research department of the Tehran University. 354 We thank two anonymous reviewers for helpful and constructive comments. 355 356 References 357 Ambraseys, N. & Melville, C., 1982. A history of Persian earthquakes. Cambridge 358 University Press. rP 359 Ambraseys, N.N. & Jackson, J.A, 1998. Faulting associated with historical and recent 360 earthquakes in the Eastern Mediterranean region. Geophys. J. Int. , 133, 390‐406. 361 Barka, A.A., 1996. Slip distribution along the North Anatolian Fault associated with the 362 large earthquakes of the period 1939 to 1967. Bull. Seismol. Soc. Am., 86 (5), 1238– 363 1254. rR ee 364 Barka, A.A., Akyüz, H.S., Altunel, E., Sunal, G., Çakir, Z., Dikbas, A., Yerli, B., Armijo, R., 365 Meyer, B., De Chabalier, J.B., Rockwell, T., Dolan, J.R., Hartleb, R., Dawson, T., 366 Christofferson, S., Tucker, A., Fumal, T., Langridge, R., Stenner, H., Lettis, W., 367 Bachhuber, J. & W. Page, 2002. The Surface Rupture and Slip Distribution of the 17 368 August 1999 Izmit Earthquake (M 7.4), North Anatolian Fault. Bull. Seismol. Soc. Am., 369 92, 43–60. iew ev hal-00681707, version 1 - 14 May 2012 349 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 12 of 28 370 Bateman, M.D., Boulter, C.H., Carr, A.S., Frederick, C.D., Peter, D. & M. Wilder, 2007. 371 Preserving the palaeoenvironmental record in Drylands: Bioturbation and its 372 significance for luminescence derived chronologies. Sedimentary Geology, 195(1‐2), 373 5‐19. 374 375 376 Daniels, J.M., 2003. Floodplain aggradation and pedogenesis in a semiarid environment. Geomorphology, 56, 225‐242. Fattahi, M., Walker, R., Hollingsworth, J., Bahroudi, A., Nazari, H., Talebian, M., Armitage, Page 13 of 28 S., Stokes, S., 2006. Holocene slip‐rate on the Sabzevar thrust fault, NE Iran, 378 determined using Optically‐stimulated Luminescence (OSL). Earth and 379 Planetary Science Letters, 245, 673‐684. 380 Fattahi, M., Walker, R.T., Khatib, M.M., Dolati, A. & A. Bahroudi, 2007. Slip‐rate estimate 381 and past earthquakes on the Doruneh fault, eastern Iran. Geophy. J. Int., 168, 382 691‐709. 383 Fattahi, M., Nazari, H., Bateman, M.D., Meyer, B., Sébrier, M., Talebian, M., Le Dortz, K., 384 Foroutan, M., Ahmadi Givi, F. & M. Ghorashi, 2010. Refining the OSL age of the 385 last earthquake on the Dheshir fault, Central Iran. Quaternary Geochronology, 5, 386 286‐292, doi:10.1016/j.quageo.2009.04.005. rP 387 Guralnik, B., Matmon, A., Avni, Y., Porat, N., Fink, D., 2011. Constraining the evolution of 388 river terraces with integrated OSL and cosmogenic nuclide data. Quaternary 389 Geochronology, 6, 22‐32. 390 Jackson, J. & D. McKenzie, 1984. Active tectonics of the Alpine‐Himalayan Belt between rR 391 ee western Turkey and Pakistan. Geophys. J. R. Astron. Soc., 77, 185‐264. 392 Le Dortz, K., Meyer, B., Sébrier, M., Nazari, H., Braucher, R., Fattahi, M., Benedetti, L., 393 Foroutan, M., Siame, L., Bourles, D., Talebian, M., Bateman, M.D. & Ghorashi, M., 394 2009. Holocene right‐slip rate determined by cosmogenic and OSL dating on the 395 Anar fault, Central Iran. Geophys. J. Int., 179, 700‐710. iew ev hal-00681707, version 1 - 14 May 2012 377 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Geophysical Journal International 396 Le Dortz, K., B. Meyer, M. Sébrier, R. Braucher, H. Nazari, L. Benedetti, M. Fattahi, D. 397 Bourles, M. Foroutan, L. Siame, A. Rashidi, & M. Bateman, 2011. Dating inset 398 terraces and offset fans along the Dehshir fault combining cosmogenic and OSL 399 methods. Geophys. J. Int., 185, 1147‐1174. 400 401 Meyer, B., Mouthereau, F., Lacombe, O. & P. Agard, 2006. Evidence of Quaternary activity along the Deshir Fault, Geophy. J. Int., 164, 192‐201. 402 Meyer, B. & K. Le Dortz, 2007. Strike‐slip kinematics in Central and Eastern Iran : 403 estimating fault slip‐rates averaged over the Holocene, Tectonics, 26, TC5009, 404 doi:10.1029/2006TC002073. Geophysical Journal International 406 Murray, A.S. & A.G. Wintle, 2000. Luminescence dating of quartz using an improved single‐aliquot regenerative‐dose protocol., Radiation Measurements, 32, 57‐73. 407 Nazari, H., M. Fattahi, B. Meyer, M. Sébrier, M. Talebian, M. Foroutan, K. Le Dortz, M. D. 408 Bateman & M. Ghorashi, 2009. First evidence for large earthquakes on the 409 Deshir Fault, Central Iran Plateau. Terra Nova, 21, 417–426. 410 411 Rittenour, T.M., 2008. Luminescence dating of fluvial deposits: applications to geomorphic, palaeoseismic and archaeological research. Boreas, 37, 613‐635. 412 R.G. Roberts, R.F. Galbraith, H. Yoshida, G.M. Laslett, J.M. Olley, 2000. Distinguishing dose 413 populations in sediment mixtures: a test of single‐grain optical dating 414 procedures using mixtures of laboratory‐dosed quartz. Radiation Measurements, 415 32 , 459‐465. rP 416 Schmidt, A., Quigley, M., Fattahi, M., Azizi, Gh., Maghsoudi, M. and H. Fazeli, 2011. 417 Holocene settlement shifts and palaeoenvironments on the Central Iranian Plateau: 418 investigating 419 DOI:10.1177/0959683610385961. linked ee systems. The Holocene, 21 (4), 583‐595, rR 420 Staubwasser, M., Sirocko, F., Grootes, P.M. and M. Segl, 2003. Climate change at the 4.2 ka 421 BP termination of the Indus valley civilization and Holocene south Asian monsoon 422 variability. Geophys. Res. Lett., 30(8), 1425, doi:10.1029/2002GL016822. iew ev hal-00681707, version 1 - 14 May 2012 405 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 14 of 28 423 Vernant, P., Nilforoushan, F., Hatzfeld, D., Abbassi, M.R., Vigny, C., Masson, F., Nankali, H., 424 Martinod, J., Ashtiani, A., Bayer, R., Tavakoli, F. & J. Chéry, 2004. Present‐day 425 crustal deformation and plate kinematics in the Middle East constrained by GPS 426 measurements in Iran and northern Oman. Geophy. J. Int., 157, 381‐398. 427 Walker, R. & M. Fattahi, 2011. A framework of Holocene and Late Pleistocene 428 environmental change in eastern Iran inferred from the dating of periods of 429 alluvial fan abandonment, river terracing, and lake deposition, Quaternary 430 Science Reviews, 30, 1256‐1271, doi:10.1016/j.quascirev.2011.03.004. 431 432 Walker, R. & J. Jackson, 2004. Active tectonics and late Cenozoic strain distribution in central and eastern Iran, Tectonics, 23, TC5010, doi:10.1029/2003TC001529. Page 15 of 28 434 Figure captions 435 Figure 1. Landsat mosaic of the Anar fault area. White squares for location of 436 photograph in Figure 2, Quickbird extract in Figure 3, and location of OSL‐2 sampling 437 site. Upper right inset locates the area within a simplified seismotectonic map of Iran. K, 438 D, A, KB, N and G, respectively for Kashan, Dehshir, Anar, Kuh Banan, Nayband, and 439 Gowk faults. 440 Figure 2. Field photograph of the Anar fault (vertical arrows, top panel) taken 441 towards the north from 31.2764°N and 55.1304°E with emphasis on a 3 m right‐lateral 442 offset rill (horizontal arrow, bottom panel). 443 Figure 3. (a) Quickbird imagery of the Anar fault trace (red arrows) centred on 444 the paleoseismological site (rectangle). Circle indicates the 8±0.5 m cumulative dextral‐ 445 offset riser described as site 1 in Le Dortz et al. (2009). (b) Topographic DGPS map (top) 446 and profiles (bottom) of the paleoseismological site. Contour interval is 5 cm (survey 447 data not tied to absolute elevation). The smooth and subdued E‐facing fault scarp is 448 indicated by a red overprint and the location of the trench by a rectangle. Violet and 449 green dots locate the DGPS data points used for the topographic profiles. (c) Field 450 photograph, looking south, of the E‐facing fault scarp. The upper part of the southern 451 trench wall is in the foreground. Three white labels that are 1‐m spaced show scale on 452 trench wall. iew ev rR ee rP hal-00681707, version 1 - 14 May 2012 433 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Geophysical Journal International 453 Figure 4. Composite photomosaic of the sediments exposed in the Anar trench 454 excavation, with top four meters (between 11 and 12 meters) east and bottom two 455 meters (between 19 and 20 meters) west of the main fault zone (a), and corresponding 456 stratigraphic log of the units with vertical positions of dated samples (b). Numbers 457 indicate units described in detail in Table 1. Three event horizons (EH‐A, ‐B and ‐C) are 458 shown as thick black lines (see text for discussion). Stars locate the positions (see exact 459 location in Figure 5) of the six samples with OSL ages given (c). 460 Figure 5. (Top) Photomosaic of southern trench wall (see Figure 3 for location). 461 (Bottom) Corresponding log, with labels marking the stratigraphic units. Faults and 462 fractures are shown with red and dashed red lines, respectively. Short black lines Geophysical Journal International represent bedding of sedimentary layers. Event horizons discussed in text are labelled 464 as EH‐A, EH‐B and EH‐C. Labelled black stars locate the OSL samples. Sample Ant‐I, 465 collected west of the fault zone at 4.1‐m‐deep within unit 2, yields an age of 13.6±1.3 ka. 466 Sample Ant‐II, collected east of the fault zone at 4.2‐m‐deep within unit 5, yields an age 467 of 12.4±0.6 ka. Sample Ant‐III, collected at 1.5‐m‐deep in a sandy lens of unit 9, yields a 468 poorly constrain ages between 11.5 and 3.3 ka. Sample Ant‐IV, collected at 1.5‐m‐deep at 469 the base of a fissure filled with eolian and runoff sand deposits within unit 12, yields an 470 age of 6.2±0.6 ka. Sample Ant‐V, collected at 1‐m‐deep within unit 11, yields an age of 471 7.1±0.7 ka. Sample Ant‐VI, collected at 0.8‐m‐deep in a sandy lens within unit 14, yields 472 an age of 6.2±0.4 ka. 473 Figure 6. Photomosaic (top) and interpretation (bottom) of the main faulted 474 zone showing evidence for the most recent earthquake (event A). Units 11 to 15, 475 affected by faulting, are warped downward, while undeformed units 16 and 17 lie 476 unconformably atop. Faults and tiny fractures are shown as red and dashed red lines, 477 respectively. ee rP 478 Figure 7. Schematic view of possible restoration of trench log showing sequence 479 of faulting and depositional phases from present‐day (step 1) to prior to the oldest 480 paleoearthquake (step 7). The trench log has been simplified for clarity. Dashed black 481 lines show inferred ground surface prior to the erosion. Step 1) is present‐day situation. 482 Step 2) is E‐facing fault scarp formed during the most recent earthquake. Step 3) is 483 restoration of the ground surface to its position prior to event A showing a vertical 484 displacement of about 60 cm created by faulting on the eastern branch. Step 4) is the 485 penultimate earthquake denoted by open fissures and cracks formed near the main 486 faulted zone. Sand‐blows also emplaced in both sides and away of the main faulted zone. 487 Step 5) is restoration of the ground surface to its position prior to event B showing 488 vertical displacement created during penultimate earthquake on the eastern and 489 western fault branches about 25 and 15 cm, respectively. Step 6) is the oldest 490 earthquake, it has produced some surficial fissures within downthrown (eastern) block 491 and a 25‐cm‐height E‐facing fault scarp. Step 7) is restoration of the ground surface to its 492 position prior to event C, showing a vertical displacement of about 25 cm. iew ev rR hal-00681707, version 1 - 14 May 2012 463 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 16 of 28 493 Figure 8. Evidence for the penultimate earthquake (event B) encountered in a 494 fissure fill and a liquefaction feature (sand‐blow) at 16 and 22 meters of the trench log, Page 17 of 28 respectively. Red lines correspond to tiny cracks postdating the penultimate earthquake. 496 (a) Photomosaic of fissure filled with eolian sands and silts and interpretative sketch. 497 The fissure is ~85‐cm‐deep and up to ~55‐cm‐wide, tapering downward through unit 9. 498 Colours denote stratigraphy as shown in Figures 4 and 5. The material at the base of the 499 fissure includes a collapsed piece from sidewall and grades into interbedded sand and 500 silt layers. The fissure is sealed by a grey, well‐stratified, surficial runoff deposit (unit 501 13). (b) Photomosaic and interpretative sketch of a sand‐blow (unit 12) made of sandy 502 material of unit 8 injected into and deforming units 9 and 11. Hydraulic fractures, 503 vertical and oblique alignments of dragged sands and gravels deform the host sediments 504 close to the liquefaction pillar. 505 Figure 9. Five well‐constrained OSL ages place bounds on the ages of the three 506 paleoearthquakes (A, B, and C) identified in the Anar trench exposure. Light grey areas 507 represent the maximum time windows for the past earthquakes, dark grey pointing the 508 minimum time window for event C. For event A, hatched domain shows the preferred 509 time interval of the earthquake (4.4±0.8 ka). Colour codes are similar to units shown in 510 Figures 4 and 5. Question mark illustrates the ambiguity due to the lack of ages in units 511 postdating the most recent earthquake (16 and 17). rR ee rP 512 ev 513 Table 1. Detailed description of the units observed in the trench excavated across the 514 Anar fault, corresponding stratigraphic column and log are shown in Figures 4 and 5, 515 respectively. 516 iew hal-00681707, version 1 - 14 May 2012 495 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Geophysical Journal International 517 Table 2. Calculated OSL ages on the Anar samples with their parameters. Ages have 518 been calculated for quartz grains with size ranging between 90 and 150 microns. 519 Geophysical Journal International Table 1 522 Unit 523 17 Light grey to cream, poorly stratified, silty sand with some angular gravels (20%, 0.5-2 cm). 524 525 16 Light cream, fairly stratified, medium consolidated, angular gravels (60%, 0.5-6 cm), with sand and silt matrix and some gypsiferous cementation. 526 527 15 Buff to light grey, well stratified, medium consolidated, angular gravels (50%, 0.5-10 cm), with sand and silt matrix and some gypsiferous cementation. 528 529 14 Grey to buff grey, thinly stratified, medium consolidated sand and angular granules (0.5-4 cm) with crossbedding. 530 531 13 Grey, well stratified, medium to well consolidated, coarse sand with angular granules and few pebbles (30%, 0.2-6 cm). The unit bottom is erosive and rests both on units 12 and 11. 532 12 Light cream to grey, sandy fissure fills and sand-blows. 533 534 11 Brown to olive cream, roughly stratified, medium to well consolidated, silts and sands with some channels of angular gravels (40%, 0.1-4 cm). The top exhibits a hard gypsiferous calcrete. 535 536 10 Cream to grey, fairly stratified, poorly consolidated, coarse sands and angular gravels (50%, 0.1-3 cm) with sandy-silty matrix and some cross-bedding. 537 538 9 Reddish brown to brownish grey, non-stratified, semi-consolidated coarse sand, with clayish silt matrix, including some channels of angular gravels (90%, 2-15 cm) with silt matrix. 539 540 8 Cream to brownish grey, fairly stratified, well consolidated, angular granules and pebbles (60%, 0.2-3 cm), with silty sand matrix, including some small cobbles and traces of sulfate calcrete within the coarser levels. 541 542 7 Grey to brownish grey, poorly stratified, poorly sorted angular pebbles (70%, 0.5-10 cm), with sandy silt matrix. 543 544 6 Cream to brown, fairly stratified, angular granules and pebbles (80%, 0.2-4 cm), with silty sand matrix, including some sparse big cobbles and silty sand lenses. 545 5 Buff to light grey, fairly stratified, loose silt and sand. 546 4 Grey to light grey, poorly stratified, well sorted, silty sands with some cross-bedding and gravelly lenses. 547 548 3 Grey, fairly stratified, well consolidated, medium to well sorted, coarse angular gravels (90%, 0.5-10 cm) with silty sand matrix, including some sparse bigger cobbles and clayish silty sand lenses. 549 2 Grey, stratified, loose, silty sand, with some cross-bedding. 550 1 Grey, poorly stratified, well consolidated, angular pebbles (>20%, 0.5-3 cm) with silty sand matrix. iew ev rR ee 551 Detailed Stratigraphic Explanation rP hal-00681707, version 1 - 14 May 2012 520 521 Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 18 of 28 Page 19 of 28 hal-00681707, version 1 - 14 May 2012 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 552 Geophysical Journal International Table 2 553 Sample Latitude °N Longitude °E Depth (m) Water (%) K (%) U (%) Th (%) Annual dose rate (Gy/ka) Ant‐I (OSL1‐b)* 31.1952 55.1534 4.1 2 0.93±0.03 1.16±0.05 4.0±0.1 Ant‐II 31.1952 55.1534 4.2 0.4 0.77±0.03 0.97±0.05 Ant‐III 31.1952 55.1534 1.7 1.1 0.97±0.01 Ant‐IV 31.1952 55.1534 1.5 0.6 Ant‐V 31.1952 55.1534 1.0 Ant‐VI (OSL1‐a)* 31.1952 55.1534 OSL‐2 31.2697 55.1337 554 Skewness CAM De (Gy) 1.64±0.05 over‐ dispersion (%) 24 0.69 22.3±1.90 3.4±0.1 1.42±0.04 27 4.09 16.1±0.90 17.5±0.47 12.4±0.6b 4.3±0.1 1.75±0.04 51 2.1 9.7±1.09 6.1±0.24 3.5±0.2b 0.94±0.01 1.44±0.05 3.7±0.1 1.78±0.14 20 0.01 11.0±0.61 0.3 0.99±0.01 1.72±0.05 4.2±0.1 1.96±0.14 25 0.35 13.5±0.90 13.9±0.70 7.1±0.7b 0.8 rP 1.23±0.05 0.6 1.21±0.01 1.60±0.05 5.9±0.1 2.20±0.05 42 0.94 11.6±1.39 13.7±0.56 6.2±0.4c 0.8 1.1 1.06±0.01 1.33±0.05 1.90±0.05 41 2.39 20.1±2.16 19.3±0.76 10.1±0.6c Fo ee rR 4.0±0.1 ev 555 a 556 b Age based on De determined using the dominant component of Finite Mixture Modeling (FMM; Roberts et al., 2000). 557 c Age previously published, now refined using Finite Mixture Modeling (FMM; Roberts et al., 2000). 558 * Parenthesis contains the sample label used by Le Dortz et al. (2009) 559 Age based on De determined using Central Age Model (CAM; Galbraith et al., 1999). iew FMM De (Gy) Age (ka) 13.6±1.3a 6.2±0.6a Geophysical Journal International iew hal-00681707, version 1 - 14 May 2012 ev rR ee rP Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 20 of 28 W Page 21 of 28 iew v Re hal-00681707, version 1 - 14 May 2012 er Pe S E r Fo 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 Geophysical Journal International N 55.15°E 3452710 3452720 a 3452700 D C 15 3452690 44 15 0 44 15 .25 43 3452670 ee rR .50 10 m 324040 Elevation (m) 1544.9 1544.7 1544.5 1544.3 1544.1 31.19 East 200 m .5 3452680 rP hal-00681707, version 1 - 14 May 2012 A c 324060 324050 324070 324080 324090 44 15 31.195 Fo Page 22Bof 28 1543.75 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 b 1544.75 31.2°N Geophysical Journal International 324100 West B VE = 20 ev 35cm A C 324040 D ie 324060 324080 324100 distance (m) w West Page 23 of 28 Calender Age (Ka) 3.5±0.2 17 16 4 15 5 9 8 A 10 ie w 6 11 Ant-II Ant-II 12.4±0.6 5 9 ev 7 7 8 Ant-III 8 Ant-V 7.1±0.7 11 6 t Ant-IV rR EH-C 12 10 ee 9 13 Ant-V Ant-VI 6.2±0.4 I 14 11 10 I 15 I 7.3±0.5 13 10.7±0.8 EH-B Ant-VI n 14 Ant-IV 6.2±0.6 EH-A 4 12 13 3 Ant-I 13.6±1.3 Stratigraphic Depth (a) rP hal-00681707, version 1 - 14 May 2012 Present-day (b) Fo 1 2 3 4 5 0m 6 7 8 9 10 11 12 13 14 15 -1m 16 17 18 19 20 21 22 23 24 -2m 25 26 27 28 29 30 31 32 33 -3m 34 35 36 37 38 39 40 41 42 43 -4m 44 45 46 47 48 49 50 51 52 -5m 53 54 55 56 57 58 59 60 (c) Geophysical Journal International 2 14 Ant-I 1 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 East hal-00681707, version 115- 14 May 2012 10 5 Meter 0 20 25 West 28 N085° 4 4 3 3 Pee 0 rR -1 4 evi Location: N31.1953 E055.1536 Elevation: 1544.5m ew EH-A 3 EH-B 2 EH-C 100 12 14 10 5 4 3 6 2 1 -1 Ant-I Ant-II 0 5 200 cm 1 0 -1 4 3 2 7 9 8 0 0 Ant-III 17 15 13 11 Ant-IV Ant-VI Page 24 of 28 1 16 Ant-V 2 Geophysical Journal International 1 For 2 10 15 20 25 28 1 0 -1 Page 25 ofEast 28 hal-00681707, version 1 - 14 May 2012 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 West Geophysical Journal International Fo rP ee Meter 12 rR 50 0 100 ev East CENTIMETERS 14 West ie 17 16 w 15 11 +13 +14 Step 1: Present-day E Page 26 of 28W Geophysical Journal International Ant-VI Ant-V Ant-I Ant-II 0 100 200 cm W 0 100 200 cm W ee rP Event A=60cm 0 rR 100 200 cm W 0 100 200 cm W w ie ev hal-00681707, version 1 - 14 May 2012 Ant-III Ant-IV Fo 1 2 3 4 5 6 7 8Step 2: Just after Event A E9 10 11 12 13 14 15 16 17 Step 3: Just before Event A 18 E 19 20 21 22 23 24 25 26 27 Step 4: Just after Event B E28 29 30 31 32 33 34 35 Step 5: Just before Event B E36 37 38 39 40 41 42 43 44 Step 6: Just after Event C E45 46 47 48 49 50 51 Step 7: Just before Event C E52 53 54 55 56 57 58 59 Event B=25cm+15cm 0 100 200 cm W 0 100 200 cm W Event C=25cm 0 100 200 cm (a) Page 27 of 28 hal-00681707, version 1 - 14 May 2012 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 (b) Geophysical Journal International Fo rP ee 0 50 15 14 13 11 CENTIMETERS rR 100 0 50 ev 15 ie 13 12 11 12 9 8 3/23 14 w 10 100 CENTIMETERS 9 8 OSL Samples Yrs X 1000 Ant-I Ant-II Event A 5.8 5.6 6.6 6.8 6.4 Event B rR 7.8 Event C ev 11.8 ie hal-00681707, version 1 - 14 May 2012 ? ee 12.3 13.0 w 12 23 3 44 5 65 76 8 97 10 8 11 12 9 13 10 14 15 11 16 12 17 18 13 19 20 14 21 15 22 23 Ant-V rP Fo 1 Ant-IV Geophysical Journal International Page 28 of 28 Ant-VI 0 14.9