Search

The Hikurangi subduction margin, New Zealand, has not produced large subduction earthquakes within the short written historic period (~180 years) and the potential of the plate interface to host large (M > 7) to great (M > 8) earthquakes and tsunamis is poorly constrained. The geological record of past subduction earthquakes offers a method for assessing the location, frequency and approximate magnitude of subduction earthquakes to underpin seismic and tsunami hazard assessments. We review evidence of Holocene coseismic coastal deformation and tsunamis at 22 locations along the margin. A consistent approach to radiocarbon age modelling is used and earthquake and tsunami evidence is ranked using a systematic assessment of the quality of age control and the certainty that the event in question is an earthquake. To identify possible subduction earthquakes, we use temporal correlation of earthquakes, combined with the type of earthquake evidence, likely primary fault source and the earthquake certainty ranking. We identify 10 past possible subduction earthquakes over the past 7000 years along the Hikurangi margin. The last subduction earthquake occurred at 520–470 years BP in the southern Hikurangi margin and the strongest evidence for a full margin rupture is at 870–815 years BP. There are no apparent persistent rupture patches, suggesting segmentation of the margin is not strong. In the southern margin, the type of geological deformation preserved generally matches that expected due to rupture of the interseismically locked portion of the subduction interface but the southern termination of past subduction ruptures remains unresolved. The pattern of geological deformation on the central margin suggests that the region of the interface that currently hosts slow slip events also undergoes rupture in large earthquakes, demonstrating different modes of slip behaviour occur on the central Hikurangi margin. Evidence for subduction earthquakes on the northern margin has not been identified because deformation signals from upper plate faults dominate the geological record. Large uncertainties remain in regard to evidence of past subduction earthquakes on the Hikurangi margin, with the greatest challenges presented by temporal correlation of earthquake evidence when working within the uncertainties of radiocarbon ages, and the presence of upper plate faults capable of producing deformation and tsunamis similar to that expected for subduction earthquakes. However, areas of priority research such as improving the paleotsunami record and integration of submarine turbidite records should produce significant advances in the future.

At Table Cape, Mahia Peninsula, North Island, New Zealand, four marine terraces have been uplifted coseismically during the past 3500 years. Detailed facies assessment of the terrace coverbed sequence coupled with identification of modern analogues on the active shore platform were used to infer the process of marine terrace formation and to estimate the timing and amount of past uplift events (earthquakes). The modern platform can be subdivided into seven depositional zones: subtidal, outer platform, intertidal sand pockets, inner platform, high-tide, mid-storm, and storm beach. Terrace coverbeds were characterised from two trenches excavated across the full width of the uplifted terrace sequence. Off-lapping packages of high tidal, mid-storm, and storm beach sediments were most common. Outer platform sediments occurred only rarely near the base of some uplifted shore platforms. Overlying the marine sediments were near-horizontal terrestrial deposits of airfall tephra (on the two highest terraces), subsoil, topsoil, rare wedges of colluvial sediment (slopewash) shed from terrace risers, and an anomalous deposit possibly emplaced by a tsunami. Fifty-one radiocarbon ages, obtained from molluscs in the marine coverbeds, showed a general pattern of seaward-younging across the coastal plain and across each terrace and a less pronounced pattern of decreasing age upward in each coverbed sequence. The distinctive stepped geomorphology of the terraces, the facies and age structure of the terrace deposits and historical earthquake causation of similar terraces elsewhere in New Zealand provided the data to invoke an earthquake-driven model for terrace formation. Marine terrace development following an uplift event involved rapid cutting of a new intertidal shore platform and generally regressive deposition of high-tide to storm beach deposits. Following further uplift, the platform became a geomorphic terrace (above marine influence) and was then mantled by terrestrial sediments. On the two highest terraces at Table Cape, airfall tephras mantling the marine coverbeds provided a minimum age for terrace uplift. The youngest radiocarbon ages from high-tide deposits high in the stratigraphy and near the seaward edge of each terrace provided the best estimates for the timing of uplift. Based on the new radiocarbon ages and the constraining airfall tephra ages, we revised the earthquake ages to 3530–3350, 1810–1730, 1560–1300 and 300–100 cal. YBP. Associated best estimates of the coseismic uplift amounts were 2.1, 1.4, 1.8, and 3.1 m respectively, once we accounted for eustatic sea level changes through the late Holocene.

Coseismic coastal deformation is often used to understand slip on offshore faults in large earthquakes but in the 2016 M W 7.8 Kaikōura earthquake multiple faults ruptured across and sub-parallel to the coastline. Along ∼110 km of coastline, a rich dataset of coastal deformation comprising airborne lidar differencing, field surveying and satellite geodesy reveals highly variable vertical displacements, ranging from −2.5 to 6.5 m. These inform a refined slip model for the Kaikōura earthquake which incorporates changes to the slip on offshore faults and inclusion of an offshore reverse crustal fault that accounts for broad, low-amplitude uplift centered on Kaikōura Peninsula. The exceptional detail afforded by differential lidar and the high variability in coastal deformation combine to form the highest-resolution and most complex record of coseismic coastal deformation yet documented. This should prompt reassessment of coastal paleoseismic records that may not have considered multi-fault ruptures and high complexity deformation fields.

Discovery and investigation of millennial-scale geological records of past large earthquakes improve understanding of earthquake frequency, recurrence behaviour, and likelihood of future rupture of major active faults. Here we present a ∼2000 year-long, seven-event earthquake record from John O'Groats wetland adjacent to the Alpine fault in New Zealand, one of the most active strike-slip faults in the world. We linked this record with the 7000 year-long, 22-event earthquake record from Hokuri Creek (20 km along strike to the north) to refine estimates of earthquake frequency and recurrence behaviour for the South Westland section of the plate boundary fault. Eight cores from John O'Groats wetland revealed a sequence that alternated between organic-dominated and clastic-dominated sediment packages. Transitions from a thick organic unit to a thick clastic unit that were sharp, involved a significant change in depositional environment, and were basin-wide, were interpreted as evidence of past surface-rupturing earthquakes. Radiocarbon dates of short-lived organic fractions either side of these transitions were modelled to provide estimates for earthquake ages. Of the seven events recognised at the John O'Groats site, three post-date the most recent event at Hokuri Creek, two match events at Hokuri Creek, and two events at John O'Groats occurred in a long interval during which the Hokuri Creek site may not have been recording earthquakes clearly. The preferred John O'Groats–Hokuri Creek earthquake record consists of 27 events since ∼6000 BC for which we calculate a mean recurrence interval of 291 ± 23 years , shorter than previously estimated for the South Westland section of the fault and shorter than the current interseismic period. The revised 50-year conditional probability of a surface-rupturing earthquake on this fault section is 29%. The coefficient of variation is estimated at 0.41. We suggest the low recurrence variability is likely to be a feature of other strike-slip plate boundary faults similar to the Alpine fault.

Fault interactions are an essential feature of all fault systems on timescales of individual earthquakes to millions of years. We examine the role of these interactions in the development of an array of normal faults within the active Taupo Rift, New Zealand. Stratigraphic horizons (0–26 ka) exposed in 30 trenches and laterally extensive topographic surfaces (∼18–340 ka) record displacements during surface-rupturing earthquakes over time intervals of up to 100’s of thousands of years. Complementary changes in displacements, displacement rates and earthquake histories between faults are observed for along-strike displacement profiles and at points on fault traces. Variations of displacement are attributed mainly to fault interactions, and decrease with the aggregation of displacements on progressively more faults and over longer time intervals. Rift-wide displacement rates are, for example, near-constant over timescales of

Determining the spatial and temporal deliveries of sediment from the Waipaoa River catchment in the post-glacial period is a major goal of integrated source-to-sink landscape evolution studies in the region. In a 2.2 km long section of the Waihuka tributary (217 km 2 ) of the Waipaoa River (2150 km 2 ) in eastern North Island, New Zealand, a sequence of at least ten fluvial terraces and abandoned meanders up to 45 m above the present river records rates and times of post-glacial downcutting. Dateable organic and tephra horizons in the terrace and meander infill stratigraphy indicate that downcutting at this location on the Waihuka tributary was dominated by a short-lived event in the early Holocene from 10–8 ka BP, when as much as half of the downcutting was accomplished in only 10–15% of post-glacial time. The large downcutting event is interpreted to record the passing of a succession of major knickpoints or a knickzone through this tributary. The initiation of knickpoint retreat in the catchment was primarily a consequence of reduction in sediment supply at the end of MIS 2. Based on the timing of rapid downcutting in the Waipaoa River mainstem, we estimate that the rate of knickpoint retreat along the Waihuka tributary was about 2 km/kyr. In the period before and after rapid downcutting in the Waihuka tributary, modest incision rates, averaging 1–2 mm/yr, probably represent the regional tectonic uplift rate.

One of the major environmental controls on sedimentary systems is tectonics, and thus in order to make quantitative explanations and predictions of the behaviour of sedimentary systems, it is important to quantify tectonic rates and mechanisms. In this study we show how a combined Holocene fluvial and marine terrace dataset can be used to place constraints on coastal uplift mechanisms at Pakarae River mouth, to the east of the Waipaoa sedimentary system. Mapping and surveying of fluvial terraces shows that the terraces can be divided into three sets, (i) PgT, a fill terrace formed during post-glacial sea level rise (~ 10000–7000 cal. yr BP), (ii) a sequence of post-7000 cal. yr BP degradation terraces, and (iii) historical (post ~ AD 1880) flood terraces. A conceptual model of terrace evolution proposes that coseismic uplift has resulted in the observed back-tilting, downstream divergence, and successive abandonment of tectonic terraces. Coastal uplift mechanisms are then examined by comparing the fluvial and marine terrace uplift distribution and the uplift-per-event from the marine terraces with those predicted by forward elastic dislocation models. The models agree with the interpretation that the terrace uplift can be explained by coseismic slip on a steeply west-dipping reverse fault located close to the shore, possibly in combination with rupture of the Hikurangi Margin subduction interface 10–24 km below. This study also provides insights into the tectonic drivers and impacts on the adjacent Waipaoa sedimentary system including: the possible role of subduction earthquakes, growth of structures on the shelf and upper slope, and the role that the interplay between sea level and tectonic controls has on sediment dispersal.

We suggest CCS risk assessment should extend beyond the primary container to address a wide range of safety, economic, social, political and engineering issues. In New Zealand, for example, past experiences of large engineering projects suggests that failure to gain public support or to allocate appropriate resources to navigate through legislative process could provide significant barriers to successful CCS implementation. We have developed a modularized and probabilistic based logic tree which encompasses the main components of CCS (capture, transport, injection and storage) and is built upon the five issues mentioned above.

The sediment flux generated by postglacial channel incision has been calculated for the 2150 km 2 , non-glacial, Waipaoa catchment located on the tectonically active Hikurangi Margin, eastern North Island, New Zealand. Sediment production both at a sub-catchment scale and for the Waipaoa catchment as a whole was calculated by first using the tensioned spline method within ARC MAP to create an approximation of the aggradational Waipaoa-1 surface (contemporaneous with the Last Glacial Maximum), and second using grid calculator functions in the GIS to subtract the modern day surface from the Waipaoa-1 surface. The Waipaoa-1 surface was mapped using stereo aerial photography, and global positioning technology fixed the position of individual terrace remnants in the landscape. The recent discovery of Kawakawa Tephra within Waipaoa-1 aggradation gravels in this catchment demonstrates that aggradation was coincidental with or began before the deposition of this 22 600 14 C-year-old tephra and, using the stratigraphic relationship of Rerewhakaaitu Tephra, the end of aggradation is dated at ca 15 000 14 C years (ca 18 000 cal. years BP). The construction of the Waipaoa-1 terrace is considered to be synchronous and broadly correlated with aggradation elsewhere in the North Island and northern South Island, indicating that aggradation ended at the same time over a wide area. Subsequent downcutting, a manifestation of base-level lowering following a switch to postglacial incision at the end of glacial-age aggradation, points to a significant Southern Hemisphere climatic warming occurring soon after ca 15 000 14 C years (ca 18 000 cal. years BP) during the Older Dryas interval. Elevation differences between the Waipaoa-1 (c.15 ka) terrace and the level of maximum channel incision (i.e. before aggradation since the turn of the 20th century) suggest about 50% of the topographic relief within headwater reaches of the Waipaoa catchment has been formed in postglacial times. The postglacial sediment flux generated by channel incision from Waipaoa catchment is of the order of 9.5 km 3 , of which ~ 6.6 km 3 is stored within the confines of the Poverty Bay floodplain. Thus, although the postglacial period represented a time of high terrigenous sediment generation and delivery, only ~ 30% of the sediment generated by channel incision from Waipaoa catchment probably reached the marine shelf and slope of the Hikurangi Margin during this time. The smaller adjacent Waimata catchment probably contributed an additional 2.6 km 3 to the same depocentre to give a total postglacial sediment contribution to the shelf and beyond of ~ 5.5 km 3 . Sediment generated by postglacial channel incision represents only ~ 25% of the total sediment yield from this landscape with ~ 75% of the estimated volume of the postglacial storage offshore probably derived from hillslope erosion processes following base-level fall at times when sediment yield from these catchments exceeded storage.

The coastal geomorphology of the northeastern Raukumara Peninsula, New Zealand, is examined with the aim of determining the mechanisms of Holocene coastal uplift. Elevation and coverbed stratigraphic data from previously interpreted coseismic marine terraces at Horoera and Waipapa indicate that, despite the surface morphology, there is no evidence that these terraces are of marine or coseismic origin. Early Holocene transgressive marine deposits at Hicks Bay indicate significant differences between the thickness of preserved intertidal infill sediments and the amount of space created by eustatic sea level rise, therefore uplift did occur during early Holocene evolution of the estuary. The palaeoecology and stratigraphy of the sequence shows no evidence of sudden land elevation changes. Beach ridge sequences at Te Araroa slope gradually toward the present day coast with no evidence of coseismic steps. The evolution of the beach ridges was probably controlled by sediment supply in the context of a background continuous uplift rate. No individual dataset uniquely resolves the uplift mechanism. However, from the integration of all evidence we conclude that Holocene coastal uplift of this region has been driven by a gradual, aseismic mechanism. An important implication of this is that tectonic uplift mechanisms do vary along the East Coast of the North Island. This contrasts with conclusions of previous studies, which have inferred Holocene coastal uplift along the length of the margin was achieved by coseismic events. This is the first global example of aseismic processes accommodating uplift at rates of >1 mm yr−1 adjacent to a subduction zone and it provides a valuable comparison to subduction zones dominated by great earthquakes.

Paleoenvironmental reconstructions of three coastal waterbodies in Wellington, New Zealand, reveal that sites were isolated from the sea within the last 7500 years through coseismic uplift and barrier growth. Evidence for coseismic uplift consists of distinct transitions in diatom assemblages representing large changes in relative sea-level or water-table level, commonly in association with sedimentological evidence for catchment disturbance or marine influx. Transitions are abrupt, laterally extensive and synchronous within each waterbody. Amount of change across transition horizons is assessed using quantitative estimates of paleosalinity and waterbody type as proxies for relative sea-level change. Seven transitions involve large paleoenvironmental changes and provide evidence for earthquakes occurring at approximately 5100, 3200 (recorded at two sites), 2300 (recorded at two sites), 1000 cal years BP and 1855 AD. Five other transitions involve smaller paleoenvironmental changes and are considered to be consistent with effects of earthquakes but do not provide independent evidence for earthquake occurrence. These smaller transitions occur at approximately 6800, 3600, 2200, 1000 (coincident with a large transition) and 500 cal years BP. The data refine ages and provide information about the extent and effects of past large earthquakes in the region. These are the first paleoecologically derived earthquake signatures for Wellington and they contribute to the sparse collection worldwide of off-fault sedimentary earthquake records for predominantly strike-slip faults.

A sequence of fluvio-estuarine sediments exposed beneath the highest Holocene marine terrace at Pakarae, North Island, New Zealand, records the early-mid Holocene infilling of the Pakarae valley. This sequence was developed on an active, coseismically uplifting coastline and provides a valuable comparison to widely used facies models for estuaries, which were developed exclusively from stable coastal settings. We describe eight sedimentary sections, distributed along a 220 m stretch of riverbank and present twelve new radiocarbon ages. Sedimentology and benthic foraminifera are used to divide the sequence into eight bio-lithofacies. These units are grouped into four paleoenvironmental facies associations: barrier, estuarine, estuary-head delta and floodplain. We compare the distribution of the Pakarae paleoenvironmental facies associations to those in models of incised valley infill sequence models and case studies of infilled valleys. These data allow us to present new contributions to the development of a facies model for the sedimentary infilling of an incised valley system that was experiencing coseismic uplift synchronous with deposition. We suggest the distinctive characteristics of such a model would include (1) part, or all, of the transgressive and lowstand sequences may now lie above modern sea level, (2) the transgressive sedimentary sequence is typically condensed relative to the coeval amount of eustatic sea level (SL) rise that occurred during that period, and (3) evidence of relative SL falls, such as transitions from estuarine to fluvial environments, despite conditions of rapid and continuous eustatic SL rise.

A correlation of fluvial terraces is presented for eight non-glacial catchments of the eastern North Island, New Zealand, within the actively uplifting Hikurangi Margin. Using a combination of loess and tephra coverbed stratigraphy, and radiocarbon and OSL dating of fluvial deposits and loess coverbeds, we demonstrate correlation of four fill terraces, T1–T4. The available age constraints suggest T1=15–30 ka, T2=31–50 ka, T3=50–70 ka, and T4=∼115 ka, but the association and temporal link with loess deposits suggest correlation with cold periods, and thus the refinement of T2=late MIS 3 (31–40 ka), T3=MIS 4 (55–70 ka), T4=MI Substage 5b (∼90 ka), 5d (∼110 ka), or MIS 6 (∼140–160 ka). The ability to correlate terraces between catchments, plus the lack of independent evidence for tectonic triggering events, indicates terraces have probably formed in response to either climate (terrestrial) or baselevel (sea level) control. Climate control is indicated by the temporal link of post-glacial incision with re-establishment of forest cover, and of LGM aggradation with limited grass and shrub cover and periglacial processes. Aggradation due to increased sediment supply under reduced vegetation is dramatically demonstrated by formation of the Taupo Pumice Alluvium terrace in response to inundation by volcanic deposits (unwelded ignimbrite) following the 1.8 ka Taupo eruption, and the response to post-settlement (

Fault growth in the upper crust is mainly achieved by the accumulation of slip during earthquakes. Large faults with cumulative displacements of kilometres experience hundreds of large magnitude earthquakes but the precise manner in which maximum earthquake size evolves over millions of years is uncertain. To bridge the gap between faulting on earthquake (e.g. 1 Ma) timescales we examine the Cape Egmont Fault in offshore New Zealand, using good quality seismic-reflection data tied to wells and seabed bathymetry. Displacements on nine horizons, including the seabed, three horizons of 14–225 ka in age and five horizons of 5.3–1.6 Ma in age, were mapped along the ca. 70 km length of the fault. These horizons record maximum displacements ranging from 6 to 2344 m and suggest that this active fault commenced its latest phase of movement about 3.7 Ma ago. Displacement profiles for all horizons and inter-horizon increments are of similar geometry, indicating an approximately constant fault length and fixed maximum displacement position during growth. Fault length was established rapidly and inherited from an underlying Cretaceous structure with fault dimensions in the overlying sequence mainly achieved by up-dip propagation. As the magnitude of earthquakes is proportional to their rupture length, the maximum size of earthquakes on the fault remained constant for at least the last 3.2 Ma, with the distribution of slip along the fault during these events approximately comparable. This contrasts with the maximum displacement rates which varied by an order of magnitude during the last 3.7 Ma (0.18–2.8 mm/yr). Decreases in displacement rates were accompanied by proportional increases in the average earthquake recurrence intervals (ca. 1.3–18 ka), and reflect regional changes in the rates of extension.

Sediment cores collected in a coastal lagoon a few kilometres east of Wairoa, northern Hawke's Bay, New Zealand, were examined using sedimentological, geochemical, palynological and micropaleontological analyses. A distinct short-lived catastrophic saltwater inundation (CSI) about 6300 years BP and possibly other minor marine incursions are preserved in the coastal estuarine to lagoonal freshwater sedimentary sequences, which have been deposited in the last 6500 years. The CSI is characterised by a gravel unit that thins landward and decreases in particle size to sand, within a sequence consisting mainly of brackish estuarine muds. Diatom assemblages indicate a marked change from the shallow brackish estuarine muds to marine gravels and sands to brackish estuarine muds. The marine influence in the gravel and sand is also shown by the presence of marine dinoflagellates and a peak in Na/Rb. Sedimentological, chemical and paleontological (in particular diatoms) evidence indicates it is a CSI. We conclude that this was a tsunami and propose the most likely propagating mechanisms. Marine influence decreases upcore and totally freshwater conditions are evident in the upper section of the cores. The geochemistry of the sediments mainly reflects the change in stratigraphy, with distinct signatures for tephra (Na, Fe, Cr), organic-rich and peat units (As, Br) and the coarse gravel-sand CSI unit (Na/Rb, Cr, Fe), but it is also indicative of changes in depositional environment. The change in chemistry (Na/Rb) in the CSI event is indicative of a saltwater influence, whereas a marked change in S content suggests a sudden change from brackish to freshwater conditions shortly after 4800 years BP. Another peak in S and Br content about 3200 years BP may indicate another temporary change to brackish conditions.

An extensive terrace (Waipaoa-1) that can be traced for about 29 km in the Waipaoa valley, eastern North Island, New Zealand, is underlain by at least 10 m of coarse, aggradational, river gravels. Terrace cover beds contain tephras erupted from central North Island volcanoes and these provide minimum ages for the underlying gravels. Tephra or tephric layers occurring in the lower cover beds were investigated at five sites using a combination of stratigraphy, mineralogy, and the major element composition of glass shards together with discriminant function analysis (DFA). The basal tephra is identified as the ca. 14,700 14C years old (ca. 17,700 cal. years B.P.) Rerewhakaaitu Tephra, erupted from Okataina Volcanic Centre. Using the stratigraphic relationship of Rerewhakaaitu Tephra, the end of aggradation is dated at ca. 15,000 14C years (ca. 18,000 cal. years). Correlation with aggradational terraces elsewhere in North Island and northern South Island indicates that aggradation ended at the same time over a wide area and confirms a climatic origin for the terraces. Subsequent downcutting was apparently rapid because Rerewhakaaitu Tephra also occurs at the base of cover beds on a ca. 15 m lower terrace. The downcutting represents a major change in river dynamics and is most likely the response to climatic change and the resultant upper catchment landscape stability.

Many landslides disrupt the coral terraces at northeast Huon Peninsula, Papua New Guinea. Each landslide is younger than the youngest terrace which it intersects and is older than the oldest coral reef or shoreline that overlies it. The ages of the raised coral reefs and terraces were established by previous studies using radiocarbon and uranium-series dating. The times of occurrence of 26 late Quaternary landslides are reported here; 7 are based on new radiocarbon dates from in situ corals or molluscs associated with landslide deposits, and the remainder are based on terrace ages that were established previously. The landslides include slumps, block slides and block flows which mostly involve failure of coral limestones of the terraces, and debris flows which originate from large landslides that cut through the coral limestones into underlying fan-delta and tuffaceous sedimentary rocks. The proportion of terrace area affected by Late Quaternary landslides ranges from 50% per km2 and depends on average gradient of the landsurface, tectonic uplift rate, and the thickness and geometry of the raised reefs and terraces. Clusters of landslides, which occurred at about 800–1250 and 6300–7800 yr BP, may have been triggered by large earthquakes which also caused metre-scale coseismic uplift. However, we did not identify landslides that could have been associated with every large coseimic uplift that has been documented in this region.

Estuarine and beach deposits in the vicinity of the present coastline at Pakarae River record the infilling of an estuary and subsequent development of a sequence of seven marine terraces during Holocene time. At the maximum of the last glaciation about 18,000 years ago the shoreline at the ancestral Pakarae River was approximately 20 km east of the present shoreline. By about 9000 years BP the sea had transgressed across most of that coastal plain to lie within a few hundred metres of the base of the present coastal hills. Seventeen radiocarbon ages from estuarine deposits record the overall rise in post-glacial sea level, but in the period c. 9500-7000 yrs BP there are reversals to the overall rising trend. Between 9500 and 8500 yrs BP there appears to have been a eustatic fall in sea level of at least 4 m. This observation is supported by data from several other localities around New Zealand. Maximum transgression occurred about 6500–7000 yrs BP when the sea reached the base of hillslopes and an extensive estuary existed behind a barrier bar. Since that time the barrier bar disappeared, probably due to stranding in an uplift event, and the coastline advanced progressively outward toward its present position. Coastal progradation (sea level regression) and subsequent erosion have occurred in association with episodic large earthquakes at about 6700, 5400, 3910, 2450, 1570, 1000 and 600 yrs BP. The present distribution of terraces has been influenced by coastal erosion, which has removed all trace of some terraces from some areas, and river erosion has modified the marine terraces near the river.

The Rangitaiki Plain, eastern Bay of Plenty, New Zealand, is the geographic expression of an active continental back-arc rift known as the Taupo Volcanic Zone. The rift is characterized by tectonic subsidence at rates of 0.4–2.0 mm/year and is bounded by faults and marginal uplift at approximately 1.0 mm/year. Since 6.5 ka, the Rangitaiki Plain prograded at least 10 km, due to volcanoclastic sediment supplied by three major rivers, to reach its present location at about 1.8 ka. At this time a linear coast was formed, cliffed to the west in marine sediments and unwelded ignimbrites and extending as far east as a greywacke headland at Whakatane. In the tectonically subsiding Rangitaiki Plain, an abundant supply of sediment overwhelms the tectonic control on coastal evolution. Further east, a wide embayment at Ohiwa remained little changed from 6.5 to ca. 2 ka, when the coastline in the plain prograded as far as the Whakatane headland and longshore drift began to provide it with a substantial volume of sediment. Barrier spits formed rapidly and sedimentation has effectively closed and partially infilled the embayment. Ohiwa estuary, despite its location in the actively uplifting margin of the continental rift, remained submerged until sufficient sediment was available to contribute to the coastline-straightening process. The Bay of Plenty coastline is presently in a near-equilibrium state.

Fault behaviour appears to be different in different tectonic provinces of New Zealand. Uplifted Holocene marine terraces along the east coast of the North Island suggest time-predictable behaviour on reverse faults of the Hikurangi subduction margin. Major dextral faults of the strike-slip fault province in central New Zealand appear to show characteristic slip behaviour. Short normal faults in the Central Volcanic Region show highly variable increments of displacement. Finally, reverse faults of the Central Otago and Northwest Nelson region may have intermittently characteristic patterns of activity. Each of these behaviour patterns has been observed elsewhere, although, where there are good data to constrain the actual dates of palaeoearthquakes, a uniform slip or coupled model best explains most situations. Temporal and spatial clustering of fault rupture and major earthquakes are increasingly being recognized. It is crucial to obtain better dates and to determine the extent of individual rupture segments in order to better characterize fault behaviour. Non-characteristic fault behaviour has serious implications for current methods of seismic hazard evaluation. Seismic hazard studies should utilize specific fault data in preference to assuming a particular fault behaviour model.

Along the 500 km long eastern coastline of the North Island, New Zealand there are, at most, seven distinct Holocene marine terraces aged 7 ka B.P. or less. The highest of the terraces is 27 m above present-day mean sea level. The coastal region is subdivided into fourteen distinct subregions based on radiocarbon ages of marine deposits overlying wave-cut shore platforms and geographic distribution of similarly-aged terraces. Holocene marine terraces are the result of uplift associated with large earthquakes (co-seismic deformation). This conclusion is based on characteristic stepped terrace morphology, clustering of ages of terrace deposits within subrogions, and the occurrence of co-seismic uplift in historic time. Differential uplift across structures and distinct age variations at subregion boundaries are also characteristic. Recurrence intervals of uplift in any one terrace sequence vary from ∼ 0.4 ka to ∼ 2.0 ka, and individual amounts of uplift vary from ∼ 1.0 m to ∼ 4.0 m. In the past ∼ 2.5 ka ages from all terraces within subrogions indicate at least 21 paleoseismic events affecting coastal areas of eastern North Island in that period. These events cluster in time, and in separate parts of eastern North Island several earthquakes occurred ∼ 0.3, 0.6,1.0,1.5, 2.1 and 2.3 ka B.P. Our data strongly support the concept of segmentation of deformation along the subduction margin. The likely cause of coastal uplift is movement on steep reverse faults (local structures rather than the subduction thrust) that propagate from or near the subduction thrust some 20–25 km below the region. Earthquakes of moment magnitude 7.3–8.0 are estimated to be associated with these fault movements.