<p><strong>Determining the location and recurrence of seismic slip at subduction zones is essential for constraining the hazard posed by great earthquakes. Although modern geodetic observations have revealed interseismically locked regions of the plate interface, it is unclear to what degree this reflects the potential for future rupture given the broad spectrum of possible slip behaviour and short instrumental record. Developing paleoearthquake records over multiple seismic cycles is therefore a powerful method of understanding spatiotemporal patterns in seismic slip behaviour at subduction margins. Conditionally stable parts of the plate interface adjacent to locked patches are identified as particularly important places to constrain long-term behaviour due to their potential to host multiple styles of interface slip.</strong></p><p>This thesis explores temporal variation in slip behaviour at subduction zones by investigating the paleoseismic record on the weakly coupled central Hikurangi margin in New Zealand. Previously, interpretation of paleoearthquake evidence in this region has been hindered by poorly constrained chronology, sparse distribution of study sites and incomplete consideration of source faults. Generating new paleoearthquake records was therefore necessary to constrain the timing and extent of large earthquakes, and to decipher whether they could represent rupture of the plate interface. Where precisely-dated tephra isochrons are available, they can provide a useful tool for accurate chronostratigraphic correlation between paleoseismic sites and improved temporal constraints within age-depth models of Holocene paleoearthquake chronologies. The age of the Waimihia tephra isochron was refined to 3574–3478 cal yr BP using integrated age modelling of 42 radiocarbon dates from proximal, distal onshore, and distal offshore locations. Using the refined isochron age alongside 68 radiocarbon dates, a new 5000-year record of four precisely-dated paleoearthquakes was generated at the Pakuratahi Valley near Napier, as a key locality for examining onshore coseismic deformation on the central Hikurangi margin. The evidence of both coseismic subsidence and uplift could have been caused by earthquakes on upper plate faults and/or the plate interface. To better distinguish between these sources and expand the paleoseismic record in space and time, a 7000- year record of seismoturbidites was developed from a suite of submarine sediment cores spanning the central margin. Integrating the onshore and offshore datasets permitted better characterization of evidence for widespread synchronous deformation across the central margin. The new interpretation of paleoseismic correlations suggests rupture of upper plate faults could generate some of the observed patterns of surface deformation, but that seismic slip on the subduction interface is likely for most earthquakes.</p><p>Temporal agreement between the central Hikurangi paleoearthquakes and subduction earthquakes recorded on the southern, locked section of the margin provides strong evidence that multi-section ruptures, spanning a distance of >300 km, may have occurred. Examining the pattern of coseismic deformation on the central margin alongside current understanding of geophysical properties highlights the weakly coupled region of the interface between slow slip zones as an important location for propagating rupture along-strike from the locked patch. Conditional stability of this region is implied as a mechanism for supporting coseismic slip on the currently creeping part of the central Hikurangi margin. These observations add to the growing literature that suggests weakly coupled parts of the plate interface adjacent to locked patches can host multiple modes of interface slip and should therefore be considered a region of significant seismic hazard, due to their ability to support great (>Mw 8) multi-section subduction earthquakes.</p>