<p>The Wairarapa Fault is a major active dextral strike-slip fault in the southern North Island of New Zealand. It is expressed at the surface as a highly segmented fault trace that occurs ~100-400 m east of the Tararua and Rimutaka Range front, where it cuts Last Glacial Maximum (LGM) aggradation gravels. Along its central section, northeast of the Waiohine River, the Wairarapa Fault progressively displaces a flight of post-LGM degradation terraces incised into the "Waiohine" surface at the top of these gravels. Detailed mapping of the variably displaced fluvial terraces at Waiohine River using a Real-Time Kinematic (RTK) Global Positioning System (GPS) and an Electronic Distance Meter (EDM) was used to produce Digital Elevation Models (DEM) from which, the displacements of terraces, risers and paleochannels could be quantified. The timing of fluvial terrace abandonment by the Waiohine River was constrained from previously published and newly obtained Optically Stimulated Luminescence (OSL) dating of post-abandonment sands and silts that mantle the Waiohine terraces. The cumulative dextral displacements of terrace risers range from 12.4 +/- 0.8 m (2 sigma) for the youngest terrace, to 101.1 +/- 3.4 m (2 sigma) for the oldest terrace (this is the Waiohine aggradation surface). The corresponding cumulative vertical displacements of the treads of these terraces range from 1.3 +/- 0.02 m (2 sigma) to 19.7 +/- 0.09 m (2 sigma). The ratio of horizontal to vertical slip for the seven Waiohine terraces averages 6.9 +/- 3.5 (2 sigma). A comparison between the paleochannel and adjacent riser displacements at the Waiohine terraces suggests that a complete riser trimming model for progressive fault displacements of flights of degradation terraces is most appropriate at this site. Under a complete trimming model, a riser begins recording displacement only after its lower bounding terrace tread is abandoned. We estimate the Late Quaternary dextral slip rate for the central part of the Wairarapa Fault by measuring the lateral offsets and abandonment ages of two terrace-riser pairs. Based on previously published OSL age data, and our new estimated displacement for the Waiohine aggradation surface, we calculate a dextral slip rate of 12.0 +/- 3.5 mm/yr (1 sigma) for the central Wairarapa Fault. The dextral displacement and maximum OSL age determined in this study for the next younger terrace require a minimum dextral slip rate of 7.2 +/- 0.8 mm/yr (1 sigma). We also determine the magnitudes of the inferred 1855 (smallest displacement) and penultimate (next-smallest displacement) single-event displacements at Waiohine terraces, as 12.4 +/- 0.8 m (2 sigma) and 9.7 +/- 1.7 m (2 sigma), respectively, implying a mean co-seismic dextral displacement of 10.6 +/- 2.6 m (2 sigma) at the Waiohine River for the last two earthquakes on the Wairarapa Fault. Previous studies suggest a northward decrease in slip rate, but are inconclusive as to the mechanism by which this decrease is accommodated. By comparing our slip rate and single-event displacement data with results from elsewhere along the fault, we infer that there has been a northward decrease in both long-term dextral slip rate and the mean size of co-seismic dextral displacements. Co-seismic dextral displacements are 10-30 % lower at Waiohine River than on the southern Wairarapa Fault (alongstrike distance of 15-20 km), and 10-50 % lower on the northern section of the fault near Mauriceville than at Waiohine River (along-strike distance of ~35 km). The decrease in Late Quaternary dextral slip rate northwards along the fault is therefore, probably a longer-term expression of such an along-strike reduction in the mean size of co-seismic strike-slip on Wairarapa Fault ruptures. This reduction may have been caused by a slip transfer off of the main fault and onto some combination of the Carterton, Masterton and Mokonui Faults, that splay northeast-ward off of that fault. The surface trace of the central section of the Wairarapa Fault is characterised by a series of left-stepping en echelon fault segments, where deformational pressure bulges have formed in the area of overlap between adjacent discontinuous strands. We mapped the fault and associated deformational surface features through a combination of field mapping and aerial photograph surveys. In addition, we used an RTK GPS to collect detailed topographic data across two particularly well defined pressure bulges, just south of the Waiohine River. Using this quantitative topographic data, we calculated the volumes of these two bulges, and the depths at which the faults bounding these bulges converge into a single fault plane. The near-surface 3D geometry of the Wairarapa Fault is defined by a three-order hierarchy of faults, where the lower two orders of faults are arranged in an en echelon array with respect to the next higher order fault. We define the first-order fault to be the single, northwest-dipping Wairarapa Fault plane that we infer to exist at depth within basement rock. The second-order (Type A) fault segments are defined to be 2-7 km long and separated by stepover widths of 250-350 m. These stepovers are where the largest (Type A) bulges have formed. Our volumetrically calculated fault convergence depth for Type A bulges are ~100-260 m, suggesting that Type A faults converge into a single Wairarapa Fault plane within basement rock, and well below the LGM Waiohine gravels. We infer that the Type A faults and bulges began forming within basement rock prior to Waiohine gravel deposition. The distinctly smaller, third-order (Type B) fault segments are 500-4000 m long and separated by fault stepover widths of 30-150 m. Between these segments, the smaller (Type B) bulges have formed. Our volumetrically calculated fault convergence depths for Type B faults are ~1-18 m, suggesting that the bulges bounding faults converge downwards at or near the base of the Waiohine gravels. The Type B bulges, therefore appear to have formed by distributed deformation of the near-surface Waiohine gravels after they had buried the pre-existing Wairarapa Fault plane in basement. The currently active Wairarapa Fault, located 100-400 m east of the range-front, we infer to be a relatively immature splay that has recently propagated upward through previously unfaulted material from an older Wairarapa Fault at depth, reflecting eastward propagation of deformation into the Wairarapa Valley in response to topographic loading of the Tararua and Rimutaka Ranges. The segmented, discontinuous characteristics of its surface fault zone suggest that the current fault strand is at an early stage of its evolution even at depth in basement, an apparent immaturity that is not simply restricted to the deformation of the surficial Waiohine gravels (e.g. Type B bulges). By reference to the results of previously published analogue models and up-scaling of these results to the dimensions of the natural setting, we qualitatively estimate a finite dextral displacement on the currently active strand of the Wairarapa Fault of between ~130 m and ~1700 m. Together with our new Late Quaternary dextral slip rate estimate for the central Wairarapa Fault, this would seem to imply an age of inception the currently active splay of the Wairarapa Fault in the Wairarapa Valley of perhaps 100-250 ka or younger. In the natural case of the Wairarapa Fault, en echelon fault segments ("R-faults") strike only 2 [degrees]-18 [degrees] (average of ~6-7 [degrees]) with respect to the average strike of the main fault in basement. This angle is much smaller than is modelled in analogue experiments of the deformation of a previously unfaulted overburden above a strikeslip to slightly oblique-slip basement fault (10 [degrees] -30 [degrees]). The width of the fault zone is also much narrower (350 m) for the Wairarapa Fault, than the scaled-up widths of the fault zones produced in these analogue models (1-2 km). These differences are interpreted to reflect the shallow fault convergence depths for the Wairarapa Fault in comparison to those created in the analogue models, which include a thick surficial "overburden" of unfaulted sand. That thick overburden allows the fault at depth to propagate upward as a wide splay-bounded fault zone at the surface. The surface structures along the Wairarapa Fault exhibit a strong plan-view asymmetry that is reflected in the consistently triangular shape of Type A and B pressure bulges and the across-strike asymmetry defined by the geometry of the faults and bulges in profile. Moreover, there is a partitioning of oblique slip components between the two different splays bounding a given pressure bulge along the Wairarapa Fault, where the northwest bounding fault strands exhibit a greater proportion of strikeslip displacement, and the southeast bounding faults exhibit a greater proportion of dip-slip displacement. This asymmetry is interpreted to be a result of the obliquity of convergence and the northwest dip of the master Wairarapa Fault plane at depth.</p>
