Paleoseismicity of the Wellington Fault at Putara: Implications for fault segmentation and earthquake timing

Executive Summary

This report focuses on reconnaissance mapping of the southern-most end of the Pahiatua section of the Wellington Fault, excavation and logging of three trenches at a site near Putara on the Dougan farm, construction of a micro-topographic map and morpho-structural map of the trench site and radiocarbon dating of fifteen organic samples from the three trenches. These data have been analysed to construct an earthquake event chronology recurrence interval and slip rate for the Dougan site. The event chronology has been compared with other paleoseismic studies along the Wellington and Mohaka Faults. The consistency of our results have been tested by analysing the slip rate and recurrence interval values derived from the Dougan site. Several different possible rupture scenarios of the Wellington Fault were investigated upon comparison of the previous and current paleoseismic studies.

Field reconnaissance mapping was undertaken along a 7 km stretch at the southern end of the 42 km long Pahiatua section in the Putara area. In addition to locating the Dougan site, several scarp locations were identified southwest of the site with small, probably single-event displacements. A spur with a large offset (44 ± 4 m) near the Dougan site was remeasured and used to estimate a minimum dextral slip rate of 3.3-4.8 mm/yr. Two other short reconnaissance trips were made into the Tararua Ranges, into the fault-aligned valleys of the Ruamahunga and Waiohine Rivers. The second trip, into Totara Flats, provided a number of samples that may yield paleo-earthquake data for the Tararua section of the fault.

The Dougan trench site:

The Dougan site (grid ref. S25/266590) was selected for trenching due to the presence of an uphill-facing shutter scarp and an associated fault-controlled swamp. A detailed GPS-RTK microtopographic map of the site was made. The Dougan site occupied a small releasing stepover in the fault trace. Three trenches were excavated at the Dougan site spanning a length of 75 m along the fault trace: Dougan-1 at the north end of the stepover basin; Dougan-2 across the high part of the shutter ridge and widest part of the peat swamp; and Dougan-3 across a small scarp at the southern end of the stepover basin.

Dougan-3 trench is described first due its simple record of faulting and lack of age control. This trench exposed a sequence of alluvial deposits from fans and streams draining from the main rangefront. This trench showed evidence for 2 and possibly 3 earthquakes. The only age control in this trench came from one sample of charcoal (3CH1; 407 ± 35 yr BP) located in a unit between the most recent faulting event and the second faulting event. The timing of the most recent event (MRE) from this trench was therefore younger than AD 1449-1632. The penultimate event faulting is structurally distinct, and according to sample CH1 must have occurred prior to AD 1449.

Dougan-1 trench exposed a sequence of peats, gravels and peaty gravels within a faulted graben, against a greywacke-cored shutter ridge at one end of the trench, and the rangefront hillslope on the other. Dougan-1 showed evidence for the 4-5 youngest, surface-faulting earthquake events during the last c. 4000 years. In some cases, east- and west- facing faults exposed in Dougan-1 can be shown to have ruptured in the same event. Seven radiocarbon samples were analysed from this trench. The MRE occurred prior to AD1840-1680, based on one young date and lack of pre-1840 (historical) faulting. To avoid confusion, the earlier event ages are best presented below, in conjunction with data from Dougan-2.
Dougan-2 trench was excavated through the middle of the swamp (basin) and exposed an extended record of stratigraphy and faulting. The basin section in this trench consists of an interbedded section of peats, silty gravels, gravelly peats, silts and sands. Some f these units are considered to be coseismic deposits, i.e. they formed as a consequence of surface faulting and strong shaking. Eight radiocarbon samples were analysed from this trench. A date from the bottom of the sequence (sample 2GZP; 6760 ± 60 yr BP) has a calibrated age of up to c. 7600 yr old. Clear evidence was found for the youngest 4-5 surface faulting events. Dates for several of the events were similar to those found from Dougan-1 trench. We also recognised evidence for up to 4 more events in the lower part of the trench, based on folding relations and the presence of co-seismic (and interseismic) deposits in the swamp stratigraphy. This part of the section is being dated in the next phase of this project.

Paleoearthquake ages at the Dougan site:

The best estimates for event timing come by comparing the records in Dougan-1 and -2 trenches that are c. 18 m apart at the northern end of the Dougan pull apart basin. The age of the most recent faulting event (MRE) from structural evidence and dating control is AD 1655-1666. The MRE is recognised in both trenches on the basis of a 0.3-0.5 m thick, unfaulted modern peat section (or soil) that overlies the highest fault terminations in the trenches. This result is rather precise but comes from a sequence of dates above and below the MRE horizon in Dougan-2 trench.

The penultimate event (Event 2) is recognised on the basis of a colluvial wedge on the shutter ridge (Dougan-2), upward-terminating faults, and rapid burial of faulted units by elastic units in both trenches. The preferred timing of the Event 2 (E2) from Dougan-1 and -2 trenches is AD 1183-1290. Event 3 (E3 is defined by upward-terminating faults and rotation (vertical displacements) on the main fault zone against the shutter ridge. E3 is poorly-dated at Dougan-2 trench, but well-defined by 3 dates in Dougan-1 trench. There, E3 is bracketed between BC 44 to AD 253. No age control currently exists for E4, though it must be older than E3 and younger than Event 5 (E5). E5 is best expressed by the deposition of a thick "co-seismic" gravelly peat in Dougan-1 trench. This association is common for this event (co-temporal with, but called Event 4 farther to the northeast). The timing of E5 covers the range BC 1517-1947.

Event ranges along the Wellington Fault:

The age ranges of paleoearthquakes at the Dougan site are generally consistent with data from six other trenches excavated along the Pahiatua section of the Wellington Fault (Death, Hughes and Ebbett sites). Variance in the different paleoseismic records may be due to real earthquake rupture timing differences, or to differences in the quality of the paleoseismic and dating records between work done in the early 1990's and now. We recognise that there are no apparent seismic gaps along the Wellington Fault, i.e. everywhere where dating can confirm it, there is evidence of a young (<550 yr old) surface faulting event along the entire length of the Wellington Fault. In addition, when we consider all of the paleoseismic data along the Wellington Fault between Long Gully and the Ebbett site, and onto the southern end of the Mohaka Fault, there is a striking correspondence of event timing information. All along this system there appears to be 4-5 major surface faulting episodes for the Wellington fault during the last c. 4000 years. Each episode may involve multiple large earthquakes that rupture segments of the fault system. We have presented a number of different rupture scenarios for the Wellington fault. However, due to the large uncertainties in event information and radiocarbon dating, it is difficult to determine a viable rupture scenario for the Wellington Fault system.

Slip Rate and Recurrence Interval:

The dextral slip rate for the Dougan site has been estimated from the offset of the rangefront edge and apparent opening of the Dougan pull apart basin (46 ± 3 m) in combination with the base of the section dated from Dougan-2 trench. This maximum dextral slip rate (c. 5.6-6.5 mm/yr) is in the same range as has been previously calculated. The average inter-event time for surface faulting from 5 events at the Dougan site is 741-903 yr. The actual inter-event times that we observe from the trench studies vary by a factor of 2. The recognition of 5 events in c. 4000 yr (as opposed to 4) means that the mean inter-event time calculated from the trenches (822 ± 81 yr) and via slip rate (760 & 280 yr) are very similar, and perhaps therefore, event recognition is complete at this site.