The North Canterbury and Marlborough regions of Aotearoa | New Zealand were severely impacted by almost 30,000 landslides triggered during the 2016 Kaikōura Earthquake. Of these landslides approximately 200 dammed rivers. In the study area near Waiau, rupture of The Humps and Leader faults (and associated ground motions) initiated at least 42 co-seismic landslides. The Leader Landslide is the largest of these landslides, with an area of approximately 600,000 m2 and a volume of 6-8 million m3. The landslide buried approximately 980 m of active Leader River bed length and dammed the river. The dam produced four lakes, with two remaining today and two having been breached by partial landslide collapse and knickpoint migration in the year following the earthquake. As of 2025, the landslide dam has not been completely breached and Lake Rebekah remains. The Leader Landslide dam presents a unique opportunity to chart the evolution of the active riverbed pre- and post-earthquake, for up to 2 km downstream of Lake Rebekah. The river’s evolutionary timeline was observed using LiDAR, satellite aerial imagery, and drone surveys from 2001 to 2024 to develop maps and topographic difference models. Key timeframes for riverbed change events were also constrained with information and dated photography gathered from previous communications with the landowners at Woodchester Station, where the landslide is located. Finally, Schmidt Hammer testing of the Pliocene-Miocene Greta Siltstone Formation was conducted to investigate the role of bedrock strength on the rate of riverbed erosion. I present the history of evolution of the Leader River, pre- and post-earthquake, and consider factors impacting riverbed morphology changes. Despite the stability of Lake Rebekah, these data show that the position and morphology of the Leader River has changed significantly in response to the landslide, with the formation of two knickpoint waterfalls up to 14 m-high, four waterbodies, and diversion of the river around the landslide toe. Evolution of the river is characterised by longer periods of stasis (e.g., months to years) punctuated by rapid changes in riverbed morphology (e.g., hours to weeks) associated with incision and aggradation. In particular, the knickpoints migrated upstream at variable spatial and temporal rates. Factors controlling the rates of processes include; rain-storm events, partial lake outburst flooding, spatial changes in Pliocene-Miocene siltstone bed induration and landowner intervention to stabilise the landslide dam. An overarching conclusion of this thesis is that landforms can develop rapidly (i.e., hours to weeks) and in the absence of historical accounts, could be interpreted to have formed over hundreds to thousands of years.
Landslides are significant hazards, especially in seismically-active mountainous regions, where shaking amplified by steep topography can result in widespread landsliding. These landslides present not only an acute hazard, but a chronic hazard that can last years-to-decades after the initial earthquake, causing recurring impacts. The Mw 7.8 Kaikōura earthquake caused more than 20,000 landslides throughout North Canterbury and resulted in significant damage to nationally significant infrastructure in the coastal transport corridor (CTC), isolating Kaikōura from the rest of New Zealand. In the years following, ongoing landsliding triggered by intense rainfall exacerbated the impacts and slowed the recovery process. However, while there is significant research on co-seismic landslides and their initial impacts in New Zealand, little research has explored the evolution of co-seismic landslides and how this hazard changes over time. This research maps landslides annually between 2013 and 2021 to evaluate the changes in pre-earthquake, co-seismic and post-earthquake rates of landsliding to determine how landslide hazard has changed over this time. In particular, the research explores how the number, area, and spatial distribution of landslides has changed since the earthquake, and whether post-earthquake mitigation works have in any way affected the long-term landslide hazard. Mapping of landslides was undertaken using open-source, medium resolution Landsat-8 and Sentinel-2 satellite imagery, with landslides identified visually and mapped as single polygons that capture both the source zone and deposit. Three study areas with differing levels of post-earthquake mitigation are compared: (i) the northern CTC, where the majority of mitigation was in the form of active debris removal; (ii) the southern CTC, where mitigation was primarily via passive protection measures; and (iii) Mount Fyffe, which has had no mitigation works since the earthquake. The results show that despite similar initial impacts during the earthquake, the rate of recovery in terms of landslide rates varies substantially across the three study areas. In Mount Fyffe, the number and area of landslides could take 45 and 22 years from 2021 respectively to return to pre-earthquake levels at the current rate. Comparatively, in the CTC, it could take just 5 years and 3-4 years from 2021 respectively. Notably, the fastest recovery in terms of landslide rates in the CTC was primarily located directly along the transport network, whereas what little recovery did occur in Mount Fyffe appeared to follow no particular pattern. Importantly, recovery rates in the northern CTC were notably higher than in the southern CTC, despite greater co-seismic impacts in the former. Combined, these results suggest the active, debris removal mitigation undertaken in the northern CTC may have had the effect of dramatically reducing the time for landslide rates to return to pre-earthquake levels. The role of slope angle and slope aspect were explored to evaluate if these observations could be driven by local differences in topography. The Mount Fyffe study area has higher slope angles than the CTC as a whole and landslides predominantly occurred on slightly steeper slopes than in the CTC. This may have contributed to the longer recovery times for landsliding in Mount Fyffe due to greater gravitational instability, however the observed variations are minor compared to the differences in recovery rates. In terms of slope aspect, landslides in Mount Fyffe preferentially occurred on north- and south-facing slopes whereas landslides in the CTC preferred the east- and south-facing slopes. The potential role of these differences in landslide recovery remains unclear but may be related to the propagation direction of the earthquake and the tracking direction of post-earthquake ex-tropical cyclones. Finally, landslides in the CTC are observed to be moving further away from the transport network and the number of landslides impacting the CTC decreased significantly since the earthquake. Nevertheless, the potential for further landslide reactivation remains. Therefore, despite the recovery in the CTC, it is clear that there is still risk of the transport network being impacted by further landsliding, at least for the next 3-5 yrs.
We examined the stratigraphy of alluvial fans formed at the steep range front of the Southern Alps at Te Taho, on the north bank of the Whataroa River in central West Coast, South Island, New Zealand. The range front coincides with the Alpine Fault, an Australian-Pacific plate boundary fault, which produces regular earthquakes. Our study of range front fans revealed aggradation at 100- to 300-year intervals. Radiocarbon ages and soil residence times (SRTs) estimated by a quantitative profile development index allowed us to elucidate the characteristics of four episodes of aggradation since 1000 CE. We postulate a repeating mode of fan behaviour (fan response cycle [FRC]) linked to earthquake cycles via earthquake-triggered landslides. FRCs are characterised by short response time (aggradation followed by incision) and a long phase when channels are entrenched and fan surfaces are stable (persistence time). Currently, the Te Taho and Whataroa River fans are in the latter phase. The four episodes of fan building we determined from an OxCal sequence model correlate to Alpine Fault earthquakes (or other subsidiary events) and support prior landscape evolution studies indicating ≥M7.5 earthquakes as the main driver of episodic sedimentation. Our findings are consistent with other historic non-earthquake events on the West Coast but indicate faster responses than other earthquake sites in New Zealand and elsewhere where rainfall and stream gradients (the basis for stream power) are lower. Judging from the thickness of fan deposits and the short response times, we conclude that pastoral farming (current land-use) on the fans and probably across much of the Whataroa River fan would be impossible for several decades after a major earthquake. The sustainability of regional tourism and agriculture is at risk, more so because of the vulnerability of the single through road in the region (State Highway 6).
The M7.8 Kaikoura Earthquake in 2016 presented a number of challenges to science agencies and institutions throughout New Zealand. The earthquake was complex, with 21 faults rupturing throughout the North Canterbury and Marlborough landscape, generating a localised seven metre tsunami and triggering thousands of landslides. With many areas isolated as a result, it presented science teams with logistical challenges as well as the need to coordinate efforts across institutional and disciplinary boundaries. Many research disciplines, from engineering and geophysics to social science, were heavily involved in the response. Coordinating these disciplines and institutions required significant effort to assist New Zealand during its most complex earthquake yet recorded. This paper explores that effort and acknowledges the successes and lessons learned by the teams involved.
