Proceedings of the 12th International INQUA meeting on paleoseismology, active tectonic and archaeoseismology

the event horizon EH2 between CW1 and CW2. The age of EH2 was bracketed between the age of FSCAN4 in CW1 and the age of FSCAN3 in CW2. According to the age model, EH2 has an age of 78 ± 14 ka. The third earthquake was interpreted from the contact between CW2 and CW3 (EH3) with a modelled age of 59 ± 17 ka. This event was sealed with the deposition of CW3. The later accumulation of CW3 was given by the graben infill above the hourglass structure, which according to the age model was deposited at 16.7±1 ka. Considering the 6.2 m along-dip separation of the original surface by the SCF, we estimated a long-term slip rate of 0.06 ± 0.02 m/ky for the SCF. In the NGF-1 trench, three seismic events were interpreted; one event is related to the aperture and infill of the older cracks affecting the gravels exposed at the bottom of the trench (Crl in Figure 3B), the second event is inferred by the occurrence of a colluvial wedge and the third is expressed by the cutting relationship between the colluvial wedge and the main fault. We inferred that the second event was an earthquake, which ruptured the topographic surface forming a fault scarp. Using the thickness of the colluvial wedge, we estimate a minimum coseismic displacement of 2.2 m that corresponded to a minimum Mw of 7.0 earthquake. The fact that the main fault cuts this colluvial wedge without strong deformation is interpreted as a fault slip progressing in small increments instead of a sudden slip related to an earthquake rupture. we could constrain the age of the three events relatives to the two available pIRIR225 ages (T11-1 and T11-2 in Fig. 3B); the first event expressed by the deepest crack infill is older than 43 ± 2 ka. The ages in CW1 are likely near the age of the second earthquake, and the third event, expressed by the cutting relationship between the main fault and CW1, is younger than 43.2 ka. Considering the 6.2 m along-dip separation of the original surface by the SCF (Fig. 2A), we estimated a long-term slip rate of 0.06 ± 0.02 m/ky for the SCF. The trench NGF-2 exposes two event horizons: the first event EH1 is the unconformity separating AGs from AD1, and the second event EH2 is the unconformity separating AD1 from AD2 in the graben infill. EH1 represents the ground surface over which AD1 was deposited due to the formation of the fault scarp. Although we did not have any numerical age near the base of AD1, we could estimate this age by considering the sedimentation rate of 0.039 m/ ky of AD2. This rate was estimated considering the time and vertical distance between samples T22-7 and T22-9 (Fig. 4). Using this sedimentation rate and the modelled age of EH2, the base of AD1 has an age of 66 ± 5.5 ka, which is the age of the first event (EH1) interpreted as palaeoearthquake (Mw 7.1). The second palaeoearthquake (M w 7.2) rotated the AD1 layers and is represented by EH2. According to the age model, this event has a modelled age of 42 ± 2.9 ka. Following EH2, AD2 was deposited sintectonically inside the graben. The deformation of AD2 is sealed by the deposition of AD3 at 14.6 ± 1.1 ka. Integrating the palaeoseismological observations of trenches NGF1 and NGF2, we interpreted that the palaeoearthquake recognised in the NGF-1 trench was the same earthquake that tilted the AD1 layers in the NGF-2 trench. By estimating the overlap between the probability density functions for both events, we obtained a better constrained age for this event of 42 ± 2 ka. Using the age of the EH1 of 66 ± 5.5 ka, we calculated a slip rate of 0.07 ± 0.01 m/ky for the NGF. The recurrence interval of M w 7.0 earthquakes is estimated using the following relation (McCalpin, 1996): RI=D/S where RI is the recurrence interval, D is the mean coseismic slip, and S is the slip rate. In the SCF the recurrence interval is 33 ± 11 ky and in the NGF the recurrence is 21 ± 4 ky. D I S C U S S I O N Our main result is that the later activity of the AFS reflects upper plate extension accommodated by Mw 7.0 ± 0.1 surface-rupturing earthquakes, which activated the entire length of the studied faults (ca. 40 km). Using the relation between magnitude and area, we conclude that the entire brittle domain was seismically active (up to 20 km in depth). The few studies regarding the temporal relationship between upper plate normal fault reactivation and large subduction earthquakes indicate that both processes are not synchronised. In the Itozawa Fault, Japan, the recurrence intervals of surface rupturing earthquakes are three to seven times longer than the interval of M>9.0 subduction earthquakes (Kimura & Tsutsumi, 2023). Similar results have been derived for the Pichilemu Fault in central Chile (Jara-Muñoz et al., 2022), where the recurrence interval of Mw 7.0 upper plate earthquakes is ten to twenty times larger than that of M>8.5 subduction earthquakes. Thus, we arrive at the same conclusion, where the studied individual faults have an extremely long recurrence interval that overcomes ca. two hundred times the recurrence of M>8.0 subduction earthquakes in northern Chile. A C K N O W L E D G E M E N T S This research was supported by the FONDECYT grant 1200170 (ANID Chile) and the FONDAP grant 15110017 (ANID CHILE).